<p>The force information collected by force sensors in a force feedback system can be affected by several issues, such as installation errors and zero drift, thus leading to discrepancies between the sensed force at the master end and the actual force in the robot’s environment. To address this problem, an algorithm is proposed to identify the initial values of the force sensor and the parameters of installation errors. By using a force sensor installed at the robot’s end effector, real-time measurements of gravity and external environmental forces can be detected. Apart from collecting force and torque data at multiple robot poses, the sensor’s installation errors, load gravity, and center of mass parameters are also calculated through a parameter identification method. These parameters are then compensated for in the force feedback system to eliminate the effects of load gravity and sensor installation errors. In 10 sets of repeated experiments, gravity compensation and inertial force compensation were applied to the force sensor. The data from the force sensor were recorded continuously in the stationary and dynamic conditions of the robot. When the robotic arm was stationary and no external force was applied, the confidence interval (CI) at a 95 % confidence level was [−0.13, 0.13]. Meanwhile, during the synchronized motion of the robotic arm with an irregular trajectory, the CI at a 95 % confidence level was [−0.26, 0.26]. The sensor readings were nearly zero, thus validating the effectiveness of the proposed method.</p>