Planetary gearbox has extensive application due to its compact structure and high transmission ratio. Nevertheless, traditional fault diagnosis methods still require external sensors, which restrict their application. As the power source, permanent magnet synchronous motor (PMSM) requires much control messages which are measured by built-in sensors to enable it to rotate at a setting speed. Researches show that there is rich information relevant to the health state of equipment. However, the relationship between fault in planetary gearbox and control messages. To solve this problem, a mathematical model of PMSM driving planetary gearbox is established. Then, to simulate sun gear fault, stiffness losses are introduced to this model and solved by ODE45 in MATLAB. Finally, an experimental platform if constructed and sun gears with different degrees of fault are tested on the platform. The results show that the established mathematical model can characterize the working state of PMSM driving planetary gearbox. And the q-axis current can be used to diagnose the fault degree of gear in planetary gearbox. The established mathematical model proposes a theoretical basis for fault diagnosis of planetary gearbox using control messages of PMSM.