With the breakthrough in the technology of wireless power transmission, wireless-powered body sensor nodes are able to harvest radio frequency (RF) energy from RF-based chargers and thus operate continuously. Rational planning of the number and positions of the chargers is an effective way to improve the charging efficiency and save deployment budget. Previous studies on RF-based charger placement mainly consider the scenario that nodes are static, or convert to the static scenario using probability statistical model. With the background of mobile body area network, this paper considers the situation that users carrying sensor nodes have specific sojourn-move behavior patterns. Based on this behavior model, charger placement optimization problem is formulated with the constraint of node's non-outage probability. Both greedy and divide-and-conquer based particle swarm optimization (D&C-PSO) approaches are proposed to solve the problem. Finally, performances of the two proposed algorithms are evaluated and compared with existing path provisioning approach through various simulations. Simulation results show that the divide-and-conquer based particle swarm optimization outperforms both greedy and path provisioning approaches in the charger placement cost while it guarantees the node's non-outage probability.