Hidden Markov Models (HMM) are the widely used modeling techniques for online handwriting recognition. This paper describes both stroke based and character based methods for Assamese handwritten character recognition using HMM classifier. In stroke based method, unique strokes that are used to write the characters are grouped and then HMM modeling is done for each of these selected class of strokes. A character can comprise of one or multiple strokes. Reference set is prepared by analyzing the different combinations of strokes and the degree of confusion between similar strokes. The stroke based method comprises of two stages. First, the stroke sequences in the test character is recognized by stroke based HMM classifier and in the second stage this sequence of strokes is compared against the entries of the reference set. The character corresponding to the matched stroke sequence in the reference set is considered as the recognized character. In character based method, each character as a whole is modeled using HMM and the classifier directly predicts the character class. Experiments were performed on 141 Assamese characters, collected from 100 native Assamese writers and it is observed that character based method gives better result than stroke based method. © 2015 IEEE.