In today's society, the rate of population aging is getting faster and faster, and the health of the elderly has attracted much attention. It has been found that many elderly physical diseases are caused by the long-term bad habits and lack of exercise in the elderly. More and more studies have shown that proper exercise can prevent and reduce the occurrence of these diseases. With the development of smart sensor technology and the application of artificial intelligence technology, smart sensors have been used to monitor various body information data of the elderly, and play an important role in studying the influence of various sports exercises on the ability of the elderly to control their body posture. This article uses the intelligent sensor system to study the influence of tennis practice on the control ability of the elderly's body posture. This article first analyzes the knowledge of sports mechanics in tennis practice through literature research, formulates correct and comfortable tennis practice methods suitable for the elderly, and uses smart sensors to track and monitor the body posture data of the elderly. Then this paper uses a particle filter-based sensor target tracking model to monitor the elderly's fall, and uses a fourth-order Butterworth filter to evaluate the elderly's body posture control ability. Then this paper uses BP neural network to recognize and analyze the body posture of the elderly and predict the posture control ability. Finally, a comparative analysis of the effect of tennis exercises on the body control ability of the elderly is carried out through the control experiment of the elderly body posture test. Experiments show that tennis exercises can significantly improve the body control ability of the elderly, especially the lateral posture control.