Computer interlocking system is the core technical equipment of railway signal, which is an important link to ensure safe and efficient operation and railway station operation. At present, China railway computer interlocking system of life cycle management approach mainly refers to the relevant provisions of the traditional relay interlocking equipment, lack of the service life for electronic equipment in the system of analysis and evaluation methods, which largely limits the service life of computer interlocking system of scientific management. This paper, starting from the architecture of the interlocking system, studies a method to evaluate the life of a computer interlocking system based on neural network, in order to realize the scientific predictive evaluation of the service life of the interlocking system. It can provide reference for the operation and management of computer interlocking system in China.