Abstract: In order to solve the problem that the access control model of the existing Hadoop cloud platform lacks dynamic mechanism, a dynamic access control model based on user behavior assessment (DACUBA) was proposed for Hadoop cloud platform. In this model, a command sequence learning(CSL) algorithm was used to extract user behavior patterns from user instruction sequence, the global pattern library was used to classify user behavior and conduct behavioral assessment of user behavior classification. Then the evaluation value was applied to the access control mechanism of Hadoop cloud platform to implement dynamic access control. The experimental results verify that DACUBA is effective, and compared with other methods, this method has higher efficiency and better stability for cloud request access control.