A health assessment method for space mechanism shafting based on fuzzy C-means clustering is proposed in the light of applying the existing health assessment methods to space mechanism shafting. First, a new parameter named characteristic frequencies’ distinctness is defined. Then, the fuzzy C-means clustering model is trained by the kurtosis and characteristic frequencies’ distinctness of fine and faulted shafts. Finally, the health index is constructed according to the membership related to the fine shafting cluster. With the proposed method, the shafts can be sorted correctly according to the health state, and the reliable health index can be obtained at high rotation speeds. The practicability and effectiveness of the approach proposed have been demonstrated.