In recent years, with the continuous progress of lunar exploration, remote robotic arms face the need for highly safe, accurate, and transparent control strategies in uncertain and unstructured environments when performing tasks such as lunar base construction and facility setup. Humanoid variable impedance control methods can ensure both safe interaction between the robotic arm and the environment and high-precision control, providing a solution to the challenges of human-robot collaboration in lunar construction. This study investigates a teleoperated variable impedance robotic arm control strategy based on human impedance parameters, projecting human variable impedance parameters onto the teleoperated robotic arm to meet the interaction requirements of lunar construction tasks. This method integrates four-channel surface electromyography signals with an upper limb muscle mechanics model to construct a real-time identification system for human end-effector stiffness. Unlike traditional measurement methods, this study addresses the generalization issue of human-like variable impedance by combining personalized human physical parameters to enhance generalization capabilities. Additionally, in remote-operated variable impedance control, force feedback and visual feedback are used to enhance information transparency during human-machine interaction and trigger natural neural reflexes in the human body to adaptively adjust impedance. Finally, based on the assembly task requirements of a lunar truss construction platform, the study validated that humanoid variable impedance remote operation exhibits superior performance distinct from traditional remote operation.