This article discusses the development of a neural network to analyze the quality of ERP systems. It uses a subset of 12 quality parameters defined in ISO 25010 as input parameters for the neural network, which is based on a multilayer perceptron architecture. Through a newly created reduction methodology, the connections in the neural network reduced, resulting in a high-accuracy neural network. Created neural network can simplify and speed up the process of analyzing the quality of ERP systems, and a web application developed for user interaction with neural network.