Nonlinear receding horizon control (NRHC) has attracted much attention in recent years. NRHC can implement various constraints; however, the numerical computation frequently become unstable due to such constraints. The algorithm by using a barrier function for inequality constraint was proposed for the problem. However, the control performance with the algorithm is not sufficient. In this paper, we clarify the problem in the RH control with inequality constraints. Then, we propose a constrained NRHC method by using artificial potential for the problem. Finally, we confirm the effectiveness of the proposed control method by computer simulation of the obstacle avoidance of the robot arm.