Abstract: Because of uncertainty factors in the burning process of solid waste，it is impossible to execute fault diagnosis by using first principle models. Principal Component Analysis (PCA) was introduced in this paper for incipient fault diagnosis of a waste solid incinerator (WSI). Based on historical data，a PCA model was built to represent its normal states. Fault detection was then realized based on two statistical variables T2 and SPE. Residual Subspace Contribution Factor graphs were then utilized to assist the fault diagnosis. The research result indicated that PCA was an effective approach to incipient fault detection and diagnosis of WSIs.