The main purpose of modern information systems (IS) is to support the procedures of controlling many business processes (BP) of an enterprise. At the same time, due to a weak formalization of BP, processes of development, reengineering of various service-oriented IS and IT services are a complex problem. To address this problem, the task to improve the method for choosing IT services that meet an assigned set of functional and non-functional limitations was stated. The specific features of the original method for choosing IT services were analyzed, its main shortcomings were identified. The model of the BP precedent was modified to establish the relationship between descriptions of the precedent, functional requirements, and the used IT services. The method for selecting IT services for the IS was improved by adding the functions of requirements analysis and searching for descriptions of IT services that partially meet the stated functional requirements. In the method, the adaptive linear associator of the mADALIN neuron was used to quantify the degree of match of the functional requirement and the description of the function of the IT service. These proposals were the basis for the improved method for selecting the IT service that best fits the set of constraints that are formed. Based on the result of the conducted research, an experimental test of the improved method for selecting an IT service to solve the problem of automation of the activities of sale force of the electronic policy OSAGO was carried out. The information technology implementing the original method for choosing IT services was compared to the improved method. It was shown that the improved method makes it possible to identify situations of a match of functional requirements of a customer and the descriptions of IT services. This makes it possible to select those IT services that match the functional requirements to a degree above the assigned minimum limit