Abstract: Aiming at the module arrival rate and user behavior path of mobile multimedia applications, this paper first proposes a user behavior path analysis model for mobile App. Secondly, combined with depth traversal, an improved maximum forward path algorithm is proposed to mine the maximum forward path that users access during each session, and finally extract the mainstream path of users, analyze the arrival rate of mobile App core module. On this basis, The proposed algorithm model is designed and implemented based on the Apache Spark Cloud Computing platform and data processing framework.