To develop an advanced gaming interface through virtual environment (VE), a well defined stochastic mathematical approach for the hand motion recognition system (HMRS) must be established. Among the various gestures of the human body, hand gestures play a major role in interactions between the human and computers. The proposed work is comprised of four main stages in the HMRS. First, the synthetic image of hand gestures is stored in the database. The second step is to reduce the feature point with well suited feature extraction algorithm. Third, trained feature points of hand gestures are tested by using 2-D Hidden Markov model (2-D HMM) approach. The last stage is to optimize the number of state sequence, in Markov model with a heuristic approach called artificial bee colony (ABC) algorithm. This tends to provide the best fitness value of the state sequence in 2-D HMM to reduce the computing complexity during the training and testing phase in the classification process. The proposed work has provides better performance measures compared with other conventional techniques.