The purpose of the study was analyzing the competition of badminton athlete. Take the athlete who played eight singles between 2013 to 2014 as the object. The study method used Many-Facet Rasch Model to set the three facets, athletes (B), raters(R) and skills(S). The calculated program was based on resistance model of both sides (the host and road team). The program instruction is Model=?, -?, ?, #, R, 0.5 The basic model of score observing is log(P_mnÿk/P_mnÿ)=B_m-B_n-R_i-S_j-aF_k. Using the estimated Unexpected Responses and Bias Report to diagnose the match skills. The results were as the following. 1> The Unexpected Responses was able to collect the excellent and abnormal performance of athletes. 2> The Bias Report was able to diagnose athletes' skills such as the observation score, the expected scores difference, comparison of skill potential, the skills pairwise and comparison and the skills comparison between athletes. As the result, using Many-Facet Rasch Model is able to Multi-diagnostic athletes' skill.