With the rise of national health as a national strategy, the health of disabled people as a special group has been increasingly concerned. Rehabilitation sports for the disabled combines rehabilitation medicine and sports, which helps the disabled to treat physical diseases and adjust mental health. In this paper, we explored the method of analyzing the rehabilitation data of disabled people’s sports based on the time series model, which aims to improve the accuracy of predicting the rehabilitation effect and optimize the rehabilitation strategy. The study adopts the ARMA time series model to analyze the rehabilitation data of the disabled, and compares it with the multiple regression model (MRM), stepwise regression model (SRM) and CAR model; at the same time, it constructs a rehabilitation pathway system that includes three dimensions: institutional guarantee, scientific guidance and talent cultivation. The results showed that the maximum residual variance of ARMA time-series model was only 0.21, which was significantly lower than the other three models (MRM:3.40, SRM:3.18, CAR:3.60); the lung capacity of the experimental group was increased from 3008.59 to 3524.51, the BMI index was decreased from 22.95 to 18.89, and all the physical fitness indicators showed significant improvement in the experimental group, and the significant improvement on all of them (P<0.05). The conclusion of the study indicated that the analysis of rehabilitation data based on the time series model has high precision, and the rehabilitation path combining the perfect guarantee system, scientific fitness guidance and complex talent training can effectively improve the physical quality of the disabled and provide scientific basis and practical guidance for the rehabilitation of disabled people's sports.
