This study aimed at investigating the performances of ATP players by using special evaluation mode. Two special points existed in this evaluation model. One was the adoption of subjective evaluation method by experts, which was different from the published literatures. The other was the use of linear algebraic method- 2-tuple fuzzy linguistic computing(FLC)model to complete efficiently performance evaluation task. This study continued to use the nine evaluation measures of the published literature and had seven senior coaches/ teachers in the expert group. The weight analysis of evaluation measures revealed that the weights of 1st serve percentage and 2nd serve points won are significantly higher than those of the other seven. Among the 11 players evaluated, Nadal had the best performance, while Lu had the worst. Finally, this study proposed overall and individual improvement suggestions for players and their coaches. In this study, a simple and efficient performance evaluation procedure has been conducted successfully, and was worth to be extended to other areas. The research results and improvement proposals obtained must be translated into actual training actions for improving player performance in future.