When abnormal event occurs, operator should diagnose the plant status to decide the response operation which is required to mitigate the effect of the event. There are over 200 event in abnormal procedures. Operator should be trained to diagnose the case of abnormal status to maintain the plant safe. However, due to variety of abnormal cases, operational mental workload is relatively high to keep knowledge for each diagnostic assessment. To reduce the mental workload, operator support system based on artificial intelligence is developed. The artificial intelligence based operator support system diagnose the plant status and suggests the abnormal event which has been occurred. Convolutional Neural Networks (CNN) have been widely and successfully applied for image classification. Also, the CNN have led breakthrough in various field not only in image classification. Therefore, the CNN have been reviewed and applied to diagnosis of abnormal status in nuclear power plant operation.