Heavy quadruped robots are subjected to uncertain impact loads during foot-to-ground contact and gait transition, which can easily lead to excessive load on the foot mechanism and structural impact damage. Therefore, a sliding mode impedance control method based on environmental parameter estimation (EPESM) was proposed to solve the problem of poor dynamic performance when using hydraulic series elastic actuators (SEA) as foot ends in unstructured environments. Based on the piston displacement transfer function of the valve controlled hydraulic cylinder, an SEA impedance control model based on the position inner loop is established, with PID serving as the basic controller. To improve the dynamic performance of SEA impedance control, a stable adaptive environment parameter estimation method based on Lyapunov’s second method is constructed to compensate for the expected SEA position using feed forward compensation. To improve the dynamic performance and adaptability of adaptive environmental parameter estimation methods at different stages of SEA work, fuzzy control methods are used to optimize the adaptive parameters in these methods. Based on the SEA state equation, a sliding mode controller and a PID controller are constructed for dynamic performance comparison and analysis. Simulation results show that under variable SEA spring stiffness and variable ambient stiffness conditions, the response speed of EPESM impedance control is significantly faster. The adjustment time can be significantly reduced from an average of 5 s to within 1 s, achieving faster expected displacement and expected contact force, while keeping the steady-state error slightly reduced and the contact force error within ±6 N. Under dynamic tracking conditions, EPESM impedance control exhibits better dynamic performance, maintaining a phase delay within 0.2 s and an amplitude error of 5.2 % for an extended period after quickly entering the tracking state.