Human motion is carried out in a certain time and space dimension, and the kinematics of human motion is a quantitative description of human motion The main measurement methods of collecting kinematic parameters by camera include plane fixed-point camera method, plane fixed-point tracking camera method, plane orbit tracking camera method, three-dimensional fixed-point camera method, three-dimensional fixed-point tracking camera method, etc. In sports image analysis and processing, global features are generally used to complete rough matching, while local features have good results in strict and fine matching. In sports image analysis, high-speed local feature extraction and matching calculation are important links. This paper presents a framework design method of sports image analysis system (Abbreviation: SIAS) based on SIFT algorithm, which can greatly reduce the amount of data and improve the efficiency of image analysis. The experimental results in the actual field show that the framework design method of SIAS based on SIFT algorithm proposed in this paper is effective and efficient.