In view of the particularity of agricultural products and the shortcomings of existing agricultural products logistics distribution system, to reduce the cost of agricultural e-commerce logistics distribution and improve customer satisfaction, in this exploration, data mining technology in the field of artificial intelligence is used. Based on the analysis of the distribution efficiency of the current mainstream agricultural products logistics distribution mode, the logistics mode is optimized and improved, and the agricultural e-commerce common delivery mode is proposed. The logistics cost and customer satisfaction under the logistics mode are tested and analyzed through genetic algorithm and MATLAB software. The results show that the timely distribution rate of three modes of self-operated distribution, third-party logistics and common delivery decreases with the increase of orders; the timely distribution rate of common delivery mode fluctuates the most, followed by the third-party logistics mode; the timely distribution rate of self-operated distribution mode is the stablest, but it is relatively low. The calculation of genetic algorithm reveals that the optimized common delivery mode of agricultural products logistics has a significant effect on oil consumption, damage cost and other aspects. By implementing the common delivery mode, the oil consumption cost, the penalty cost, the refrigeration cost, and the damage cost are optimized by 26.7%, 31.7%, 30.3% and 19.6%, respectively. Moreover, 95% of customers are totally satisfied. In this exploration, the data mining technology is used to optimize the agricultural products distribution mode, which makes a beneficial exploration for solving the logistics bottleneck problem in the e-commerce environment, enriches the research of e-commerce logistics distribution system, and provides reference for promoting the development of agricultural products e-commerce logistics system and improving the efficiency of agricultural products logistics distribution.