Unmanned Aerial Vehicle (UAV) mounted base stations have been widely used to enhance the existing terrestrial communication infrastructures. In this paper, we investigate the problem of deploying Wireless-Powered UAV base stations (UAV-BSs) for maximizing Throughput (WPUT), that is, deploying a specified number of UAV base stations, which can harvest power from Wireless Charging Towers (WCTs), to maximize the total throughput for all users under the constraint of providing communication service continuously with the wireless power. To address the problem, we first approximate the nonlinear charging power as a piecewise constant function while bounding the approximation error. Then the entire region is divided into different charging subareas for the UAVs. Furthermore, we discretize the subareas by another piecewise constant function to approximate the communication rate with error bound. Finally, the problem is transformed into a 0–1 integer programming problem. We apply the primal-dual technique to deal with the problem after LP-relaxation and achieve a $\frac{1}{(1+\epsilon)^{2}}$ -approximation algorithm. Simulation results show that our proposed algorithm can outperform comparison algorithms by 36.92% on average.