While evaluating a martial arts athlete’s postural balance is crucial, there is a concern with providing reliable results. When it comes to evaluating martial arts players’ posture, the conventional neural network algorithm fails miserably. Thus, this study examines the current system for evaluating Wushu players’ postural balance and suggests an alternative system based on Higuchi’s research. Before reducing interference factors in postural balance assessment, the influencing elements are located using perceptron theory. Indicators are then split according to the needs of postural balance assessment. The next step is to design a Higuchi postural balancing evaluation method based on perceptron theory. The outcomes of this evaluation will be thoroughly examined. As far as accuracy and time spent considering factors impacting postural balance assessments go, the MATLAB simulation results reveal that Higuchi outperforms the conventional neural network approach under certain evaluation criteria.