As is well known, the abnormal sound detection is usually very difficult. In this paper, we present a new diagnostic method for abnormal sound using the Cellular Neural Network (CNN). The procedure of our method is 1) calculating the autoregressive model (AR model) coefficients from the abnormal sound by using the maximum entropy method ; 2) constructing the CNN whose memory patterns become standard abnormal sound patterns ; 3) making the coefficients obtained as an initial pattern and recalling one from the memory patterns, and then obtaining a diagnosis result. By using our method, the influence of the noise occurred from other normal parts can be avoided. Therefore the abnormal sound can be diagnosed by using CNN even when a man can not diagnoses the abnormal sound conventionally. The results demonstrate the advantages of our approach.