The invention relates to a human body motion capture data-based humanoid robot gait planning method. The method includes the following steps that: human body gait modes are extracted, angle curves of the six degrees of freedom of a single leg are obtained, wherein the human body gait modes include hip joint roll, hip joint pitch, hip joint yaw, knee joint pitch, ankle joint pitch and ankle joint roll; format conversion and mathematical optimization are performed on the human body gait mode data, so that the data can be applied to a human simulating robot; and a ZMP control algorithm is introduced to finely adjust the joint angles of a humanoid robot to improve the robustness of the humanoid robot. With the human body motion capture data-based humanoid robot gait planning method of the invention adopted, the gait planning difficulty of the humanoid robot can be simplified, the humanoid property and stability of the gait of the humanoid robot can be improved, and a foundation can be provided for the complex action planning of the humanoid robot. Compared with a traditional humanoid robot gait planning method, the human body motion capture data-based humanoid robot gait planning method has outstanding advantages and has a bright prospect.