As the traditional partial least squares (PLS) method has the problem of tedious steps in the process modeling, it is difficult to obtain the online performance evaluation model of complex process.The autoregressive projection to latent structures (AR?PLS) algorithm was used to establish the predictive model for the related process data.An online performance evaluation model was established by the performance levels of the output data which was divided by the fuzzy C?means clustering algorithm. The simulation results show that compared with the traditional partial least squares method, the modeling process of this method is simpler and the computational complexity is reduced, it has certain application value.