The ecological restoration of islands, or island restoration, is the application of the principles of ecological restoration to islands and island groups. Islands, due to their isolation, are home to many of the world's endemic species, as well as important breeding grounds for seabirds and some marine mammals. Their ecosystems are also very vulnerable to human disturbance and particularly to introduced species, due to their small size. Island groups, such as New Zealand and Hawaii, have undergone substantial extinctions and losses of habitat. Since the 1950s several organisations and government agencies around the world have worked to restore islands to their original states; New Zealand has used them to hold natural populations of species that would otherwise be unable to survive in the wild. The principal components of island restoration are the removal of introduced species and the reintroduction of native species.

Isolated islands have been known to have greater levels of endemism since the 1970s when the theory of island biogeography, formulated by Robert MacArthur and E.O. Wilson, was developed. This higher occurrence of endemism is because isolation limits immigration of new species to the island, allowing new species to evolve separately from others on the mainland. For example, 71% of New Zealand's bird species (prior to human arrival) were endemic. As well as displaying greater levels of endemism, island species have characteristics that make them particularly vulnerable to human disturbance.

Many island species evolved on small islands, or even restricted habitats on small islands. Small populations are vulnerable to even modest hunting, and restricted habitats are vulnerable to loss or modification of said habitat. More importantly, island species are often ecologically naive, that is they have not evolved alongside a predator, or have lost appropriate behavioural responses to predators. This often resulted in flightlessness, or unusual levels of tameness. This made many species susceptible to hunting (it is thought, for example, that moas were hunted to extinction in a few short generations) and to predation by introduced species. Some, such as the dodo, are thought to have become extinct because of the pressure of both humans and introduced animals. One estimate of birds in the Pacific islands puts the extinctions at 2000 species. Between 40 and 50% of the bird species of New Zealand have become extinct since 200 AD.

The field of island restoration is usually credited with having been started in New Zealand in the 1960s, but other smaller projects, such as the restoration of Nonsuch Island in Bermuda (which began in 1962) have been going on for almost as long. Nevertheless, the program undertaken by the Department of Conservation (DOC) is one of the largest in the world. It began on Cuvier Island, where ecologists removed stock, goats, feral cats and finally, in 1993, Pacific rats. The success of the project resulted in similar projects around New Zealand. The advantages to the DOC were considerable; in addition to protecting species endemic to smaller islands, like the magenta petrel, islands near the mainland, once restored, could act as habitat for species of birds that were unable to survive on the mainland. Species like the takahē, where the remaining wild population was at considerable risk from feral cats and dogs, could be moved to these islands to safeguard the species.

One important aspect of island restoration is the removal of invasive alien species. Since these species are most often the reason that native fauna and flora is threatened, their removal is essential to the restoration project. From 1673 until 2009, 786 successful invasive vertebrate eradication have been recorded and in the last few decades the frequency of eradications and the size of islands from which invasive vertebrates have been eradicated has increased. A definitive list of past island restoration efforts exists as the Database of Island Invasive Species Eradications. In addition a list of the current invasive species present on the world's islands exists as the Threatened Island Database

Islands are particularly suitable for restoration as once cleared of an introduced species they can be kept cleared of these species by virtue of being an island. Species removal is intensive and expensive, and methods used must be carefully chosen as to not create too much impact on non-target species. Feral cats, goats and three species of rats are among the most damaging species introduced to islands (Moors & Atkinson 1984). The differences in size, lifestyle and behaviour preclude the use of the same techniques for all of them, but with many species a range of techniques needs to be used in order to ensure success. Larger animals, such as goats and pigs, can be effectively hunted; in the case of Round Island, in Mauritius, all the goats were eliminated by a single marksman. On larger islands ecologists use a Judas goat, where a radio collared goat is released into the wild. This goat is then followed and groups it joins are removed. To remove cats a combination of techniques is needed: hunting, trapping and poisoning. Cats are more difficult to hunt than goats and pigs, requiring the use of experienced hunters and night hunting.

Trapping is ineffective for rats, given their sheer numbers, and the only method that works is poisoning, which can be delivered into the field by broadcasting (by hand or from the air) or by the maintenance of bait stations. This method has been employed around the world, in the Falkland Islands, in the tropical Pacific, and off New Zealand, where over 40 islands have been cleared. This method is not without problems, especially if the rats share the island with other, native species of rodent that might take the bait as well, as has happened on Anacapa Island in the Channel Islands and Rat Island (Hawadax) in the Aleutian archipelago. In the Pacific poison intended for rats was taken by land crabs instead; the crabs were not affected by the poison but frustrated attempts to clear the rats.

The removal of invasive weeds is, in most cases, more difficult than removing animal species. One such eradication was that of sandbur, Cencrus echinatus, an introduced grass on Laysan. The grass, introduced to Laysan around 1961, had taken over 30% of the island by 1991, displaced the native bunchgrass, and reduced the breeding habitat of two endemic threatened species, the Laysan duck and Laysan finch, as well as those of several seabirds. The removal took ten years, with controlled spraying for the first year, then individual removal of plants, then, when few plants were being found, sifting of the sands around plants to remove seeds. The cost of the eradication program was $150,000 per year.