This paper investigates a fog-assisted wireless energy harvesting network, where the UAV acts as the mobile wireless energy source and the fog server to charge and provide computation service to the sensors simultaneously. With the harvested energy, the sensors complete their computation tasks locally or offload them to the UAV. For such a system, a total energy consumption minimization problem for the UAV by jointly optimizing the UAV's flying trajectory is formulated, the task offloading and CPU frequency subject to the tasks computing requirements and the energy harvesting requirements being satisfied. Since the problem is non-convex and with no known solution, an efficient solution method is designed on the basis of Successive convex approximation (SCA) method. Simulation results show that the UAV energy consumption can be greatly reduced by using our proposed design, and the trajectory plays a dominant factor on the energy consumption of the UAV. Moreover, the longer the given time, the longer the trajectory length of the UAV. Additionally, with the increasing of the sensors' energy harvesting threshold or the decreasing of the energy conversion efficiency, the trajectory shifts toward the sensors more obviously. Compared with the uniform distribution of sensors, when the sensors are distributed concentrated, the UAV should fly closer to the sensors.