In today's rapid industrial development, the application of industrial robots is becoming increasingly extensive, where the industrial robot needs diversified designs to adapt to rich application scenarios and different utilization conditions. How to design products quickly and reasonable according to market demand has become an urgent problem in the development of industrial robots. The traditional design method of robots is mainly in accordance with experience, or functional analysis and module division of industrial robots which needs a lot of time cost and resources. In order to improve the design efficiency of industrial robots, this paper proposes a design method based on association rules through knowledge mining in which the robot structural feature parameters are accumulated and utilized. In this process, The K-Means clustering method is used to discretize the data through the Euclidean metric between the feature parameters, and then the association relationships are mined by using the Apriori algorithm, then the association rules are summarized according to the physical meaning of the structural feature parameters. The method proposed in this paper provides a scientific basis for the rapid design of robots and improves the rationality of robot design.