Tennis sports technology analysis is an evaluation of tennis sports technology from multiple perspectives, such as level and content, and is an important means of movement instruction for students in physical education. The application of intelligent technology for tennis sports technology analysis can reduce the workload of teachers and assist in tennis teaching. In this paper, multiple sensor nodes and the host are used to construct a wearable sensing network and realize the network communication connection. A support vector machine algorithm based on particle swarm optimization is proposed, and parameter optimization is performed on kernel function parameters and penalty factors to ensure the efficiency and accuracy of action recognition. Then a smaller number of action words are utilized to represent the original human motion sequence, and the use of angular features to represent the gestures of human motion is proposed to form a sequence of angular features. Finally, using the key gesture extraction and hierarchical clustering method, action words are extracted, and the generation of action text is realized by replacing each frame of data in the original action sequence with the action word that is closest to it. The accuracy of action recognition of this paper’s algorithm is 7.5% higher than that of the decision tree. The Spearman correlation coefficient between the quantization method of this paper and the traditional quantization method is 0.844, which shows a very strong correlation. Therefore, the overall fit between the performance rankings measured by this paper’s method and those measured by the teachers is high, proving the effectiveness of this paper’s quantization scheme in analyzing tennis sports techniques in physical education.