To improve monitoring performance, the traditional principal component analysis (PCA) based process monitoring approach has been extended to its probabilistic counterpart. However, its ability is limited in linear processes. This paper proposes a nonlinear probabilistic method for monitoring nonlinear processes, which is based on generative topographic mapping (GTM). Similar to traditional methods, the monitoring statistic and its corresponding fault diagnosis approach have both been developed. Two case studies are provided to evaluate the feasibility and efficiency of the proposed method.