A skill-rating system is a system that evaluates players’ skills and predict the winning probability based on competition records and relevant information. Take Association of Tennis Professionals as an example. Player’s performance is often affected by the surfaces. In recent years, some researchers have proposed models that take surfaces into consideration for tennis singles. Research results show that models incorporating surfaces can more accurately predict the probability of winning. The present thesis aims to extend this model to tennis doubles matches. The experimental results show that the inclusion of surfaces in doubles matches can improve the accuracy of prediction.