People are in the intellectual age. The use of technologies such as smart search engines, machine translation, fingerprint analysis, facial scanning, and self-driving makes people’s social and work lives happier. The intersection of artificial intelligence and other projects has also become the focus of research in the new era. This article uses comparative experimental methods and sampling methods, puts forward the evaluation method of association rules in the recommendation of sports goods, and compares the selection of sensors used by the five types of sports companies in the sports experiments. In the experiment part, according to the minimum confidence level, the two rules are met. Minimum confidence is a threshold defined by users or experts to measure confidence, which represents the lowest reliability of association rules. The conclusion is that 30% of users have bought running shoes and hiking shoes at the same time and 66.6% of users have also purchased sports suits; there are also sports suits purchased at the same time. All mountain bike users have bought sports gloves. The weight of Freescale is 0.2530, the weight of Bosch is 0.4457, the weight of ST is 0.0946, the weight of VTI is 0.0953, and the weight of Konix is 0.1114. The experimental results show that the selection result of the three-axis acceleration sensor is the SMB380 model of Bosch.