This paper proposes an optimal admittance controller for safe robot-environment interaction. The aim of the approach is to control the position of the robot end effector reactively according to force feedback. The sensors used are simple and provide only limited and immediate information, nonetheless they allow to reactively correct the applied force to guarantee safe robot-environment interaction. The method considers a dynamic optimization problem in the context of admittance control, which is solved via the standard gradient flow approach. As a remarkable feature, the solution of the optimization problem yields a free of dynamical robot model admittance controller. A stability proof is given while experimental results show the effectiveness of our approach.