Cloud computing has advanced over the years as utility computing, application service provision and software as a service model. Cloud computing has competency of remodeling the entire architecture of Information Technology; however, not much has been done to overcome the threats to this type of computing model related to intruder utilizing resources of registered cloud user. This paper proposes a dynamic user profiling system which tries to overcome the security setbacks of cloud computing concerning masquerading. The User Profiling System monitors the user's behavior looking for divergence of behavior from normal thus improving cloud security by taking proactive and reactive measures on encountering atypical behavior. It analyzes the behavior of the users using soft computing technique of neural networks and fuzzy logic and identifies the malicious users of the cloud computing model.