Cyber-physical systems (CPSs) have been applied to water distribution systems for the efficient system operation and maintenance as an application field of fourth industrial revolution. Although CPSs technology is an efficient technology based on information communication technology, it is vulnerable to cyber-attack due to information disturbance of sensor and communication server. These cyber-attack causes serious problems in the operation of the water distribution systems such as decreases of water supply, water pollution and damage to physical systems. Therefore, a few studies were performed to organize the various cyber-attack scenarios and cyber-attack detection algorithms were developed. In this study, an artificial neural network model is applied for the detection of cyber-attack by varying the number of neurons. The developed detection technique can be contributed to the establishment of reliable the water distribution systems in real operation.