Diagnosis of condenser abnormality is very important for turbine generator reliability. This paper presents a novel approach for condenser fault diagnosis based on kernel principle component analysis (KPCA) and probabilistic neural network (PNN). KPCA is applied to PNN for feature extraction. It firstly maps data from the original input space into high dimensional feature space via nonlinear kernel function and then extract optimal feature vector as the inputs of PNN to solve condenser fault classification problems. A global optimizer, particle swarm optimizer (PSO), is employed to optimize the parameters of PNN to improve fault classification accuracy. The experimental results show that the proposed approach has a better ability in terms of diagnosis accuracy and computational efficiency compared with a number of popular fault diagnosis techniques.