The face is a Simple multidimensional visual model and developing a computational model for face recognition is difficult. Hidden Markov Model has been widely used in various fields like face recognition, speech recognition etc... In this, a method of face recognition based on Hidden Markov Model is proposed. The face patterns are divided into several small-scale states based on emission and transition states they are combined to obtain the recognition result. In particular, the proposed method achieved 96.6% recognition accuracy for 400 patterns of 40 subjects. ORL database is used for experimental results (40 Individuals or subjects) i.e. 10 patterns in each subject. Results are compared with proposed approach of face recognition based on PCA. Experimental results show that the performance of HMM based face-recognition method is better than those of the PCA. Future enhancement of the work is the extraction of features based on similarities using Fuzzy C-Means Clustering and classification with Neural Networks.