The purpose of this study was to use Rasch measurement to analyze the skills of women doubles in table tennis. The subjects for the women's doubles group were the players named Yi-Hau Huang and Yun-Feng Lu who were the No.16 of world-ranking women's doubles on May 2008. The method was the skills analysis of the video of 10 games in women doubles. The skills were divided into three stages, which included serves-stage, receive-stage and rally-stage. The categories (F(subscript k)) of stage were 0, 1, 2, 3 and 4 based on wins and losses. This study used many-facet Rasch model to analyze four facets: home players (β(subscript n)), guest players (β'(subscript m)), raters (R(subscript m)), and skills (R(subscript j)). The model was log(P(subscript nmijk))=β(subscript n)-β'(subscript m)-R(subscript i)-S(subscript j)-F(subscript k). The results were: 1. The rating of five raters were inconsistent, by excluding one of raters, ratings had consistency. 2. The facet of overall players had good reliability, validity and representation. 3. The abilities of overall players and serves-stage for Lu were better than Huang, but the abilities of receive-stage and rally-stage for Huang were better than Lu. The conclusions of the study were as follows: the Rasch measurement fitted for skill analysis in women doubles table tennis. For the doubles team of Huang and Lu serves-stage ability for three stages was better, receive-stage was medium ability, and rally-stage ability was obviously poor.