Abstract: To predict and improve the human performance in spatial segmentation Chinese handwriting recognizer, a mathematic performance model was set up based on the Fitts’ law and mathematical expectation of the discrete variables. The model was verified by the experiment that the model could account for the variance of human performance data in the experiment very well (R square > 0.85) after the subjects experienced 8 times (8*24 characters) of training. The model can be applied into (1) the prediction of the task completion time when users are interacting with a specific spatial segmentation handwriting recognizer in copying task; (2) the choice of the optimal parameters in the recognizer; It was also found by the model that when the other parameters in the handwriting recognizer are constants, the reduction of user’s handwriting time per character will play a more important role in determining human performance than the reduction of recognition time

Key words: mathematics psychological modeling, Chinese handwriting recognizer, human performance modeling, spatial segmentation