Abstract: This paper designs a composite framework observer combining a general sliding mode observer based on a hybrid nonsingular terminal sliding mode surface with a Luenberger observer, which is used for parameter identification of surface-mounted permanent magnet synchronous motor. First, a comprehensive observer framework is employed to identify the mechanical parameters of the permanent magnet synchronous motor in real time, constructing a composite observation framework composed of an extended sliding mode observer （ESMO） and a Luenberger observer. When designing the ESMO, an algorithm for identifying mechanical parameters is developed based on the extended state equation established by the error signals of the mechanical parameters. This results in an observer structure that achieves a filtering effect, thus avoiding the need for additional filters. After obtaining two mechanical parameters using the ESMO, the load torque is identified and solved. However, the identified parameters still have issues of reliability uncertainty and coupling between different parameters. To address this, a new Luenberger observer is introduced in the composite observation architecture to perform decoupling and independent identification. The experiments in this paper are conducted on the MATLAB/Simulink platform, and the accuracy of the proposed method has been verified.