The grammar of the Marathi language shares similarities with other modern Indo-Aryan languages such as Odia, Gujarati or Punjabi. The first modern book exclusively about the grammar of Marathi was printed in 1805 by Willam Carey.

The principal word order in Marathi is SOV (subject–object–verb). Nouns inflect for gender (masculine, feminine, neuter), number (singular, plural), and case. Marathi preserves the neuter gender found in Sanskrit, a feature further distinguishing it from many Indo-Aryan languages. Typically, Marathi adjectives do not inflect unless they end in an आ (/aː/) vowel, in which case they inflect for gender and number. Marathi verbs inflect for tense (past, present, future). Verbs can agree with their subjects, yielding an active voice construction, or with their objects, yielding a passive voice construction. A third type of voice, not found in English for example, is produced when the verb agrees with neither subject nor object. Affixation is largely suffixal in the language and postpositions are attested. An unusual feature of Marathi, as compared to other Indo-European languages, is that it displays the inclusive and exclusive we feature, that is common to the Dravidian languages, Rajasthani, and Gujarati.

The contemporary grammatical rules described by Maharashtra Sahitya Parishad and endorsed by the Government of Maharashtra are supposed to take precedence in standard written Marathi. These rules are described in Marathi Grammar, written by M. R. Walimbe. The book is widely referred to students in schools and colleges.

Traditions of Marathi linguistics and above mentioned rules give special status to ‘tatsama’ (तत्सम) words borrowed from the Sanskrit language. This special status expects the rules for ‘tatsama’ words be followed as of Sanskrit grammar.^{[citation needed]}

Marathi words can be classified in any of the following parts of speech:

Nouns are primarily divided into three categories – proper nouns (विशेषनाम, visheshnāma), common nouns (सामान्यनाम, samānyanāma), and abstract nouns (भाववाचकनाम, bhāvvāchaknāma) – that are identical in definition to their counterparts in other languages (such as English), and are inflected for gender, number and case. They are also often categorized based on their ending vowel, which is especially useful in studying their inflection – those ending in the schwa (or inherent vowel) a (अ) are termed akārānta (अकारान्त), those ending in the vowel ā (आ) are termed ākārānta (आकारान्त), those ending in the vowel ī (ई) are termed īkārānta (ईकारान्त), and so on.

There are three genders in Marathi: masculine, feminine, and neuter. Some other modern Indo-European languages have lost these genders, completely, as in English and Persian, or in part, with either neuter and common gender (merging masculine and feminine), as in some Northern Germanic languages, or feminine and masculine (absorbing neuter), as in almost all Romance languages.

While there exist no concrete rules for determining the gender of a given noun, certain observations do help speakers in that regard: masculine nouns can only be akārānta or ākārānta, while neuter nouns can only be akārānta, īkārānta, ukārānta (उकारान्त, ending in u), or ekārānt (एकारान्त, ending in e).