This paper proposes a new method for controlling robotic contact force based on an online environment stiffness admittance controller on curved and complex surfaces. A second-order system models the robot’s interaction with the environment, while a position-based admittance controller adjusts the reference force. A proposed method obtains the desired force by estimating the online environment stiffness based on the robot’s stiffness and combining the damped force. An exponential stability theorem was utilized to check the stability of this controller. The simulations were conducted to determine the efficiency of this method using different robot stiffnesses. Accurate positions were found at robot stiffnesses of (0.6) and (0.8) N/m and environment stiffnesses of (8500) and (9000) N/m for both surfaces at a force of 50 N. Moreover, polishing experiments were applied on the surfaces based on the simulation results. The contact force fluctuations did not exceed ±1.25±1.3μμμμ and ±1.25±1.3μμμμ N out of 10 N for both cases. Furthermore, roughness values were reduced from ranges of 40 to − 20 and 20 to − 20 ±1.25±1.3μμμμm to ranges of 3.5 to − 3 and 5 to − 5.75 ±1.25±1.3μμμμm for a vertical and horizontal line of the curved surface, respectively. Similarly, roughness values decreased from ranges of 180 to − 60 and 0 to − 66 ±1.25±1.3μμμμm to ranges of 2.5 to − 3 and 3.75 to − 2.3 ±1.25±1.3μμμμm for the vertical and horizontal lines for the complex surface, respectively. The approach was able to track desired force on these surfaces, which is quite challenge.