Abstract: Resource allocation methods and technology have always been a hot issue in the field of cloud computing. The existing solutions for resource allocation have not considered the actual requirement of the users so far. Through introducing the concept: utility, this paper proposes a description model for the user’s utility in the cloud environment, which quantifies the user’ satisfaction about the time and cost of the task in the cloud environment. Considering the randomness of the arrival time and the type of the tasks, this paper proposes an optimization model of task scheduling based on the theory of linear programming, using the basic concepts of user utility. This model takes the total utility value of the tasks completion as a goal, and takes the user tasks’ expected time, cost and parallel speed-up ratio as the constraint condition. It can describe the randomness of the user’s tasks, and choose the fittest resources which maximize the needs of every user while keeping the interests of other users. Finally, the simulation results verify the user utility optimization model of this paper.