Abstract: A fuzzy prediction algorithm arising from conventional hybrid position/force control model is developed to deal with a force-controlled execution of compliant robot tasks in an unknown environment. According to the known contacting path of the manipulator's end effector and its motional direction, the algorithm makes use of the three prediction factors (satisfiability, curvature adaptability and reference proportion) to predict and readjust the desired trajectory given in hybrid position/force control model when sampling in future. The last prediction shall be re-evaluated first once a prediction starts and, at the same time, the environmental change in curvature and stiffness should be taken into account. The algorithm is also available to be applied to impedance control model. The simulation results verify that the proposed fuzzy prediction will result in better force control in different control models and has high adaptability to unknown environment in changing.