In the paper we deal with force control of redundant robots. Globally stable hybrid force control law is used to control force and trajectory in the operational space. Redundancy is solved using gradient optimization technique with modification, that enables use of dynamically consistent pseudo-inverse. The desired null space motion is controlled with a null space velocity controller. Hierarhical task decomposition is used to combine different optimization criteria. The performance of the proposed control law is shown in simulation of 4 DOF force controlled planar redundant robot with obstacle avoidance and torque optimization.