The role of the Internet in people’s daily lives, the impact of social networks on the formation of public opinion, the spread of mobile communications, the collection of personal information in electronic information systems in the e-government environment made the problem of “behavior analysis” even more relevant. In order to improve the efficiency of the public administration process during the formation of the information society, one of the most important tasks to be performed by the government organizations is the correct assessment and prediction of citizens’ behavior and making the right decisions. The main goal of the intellectual analysis of behavior is to understand the logic of the activities of individuals and social groups. This article studies the international practice in intellectual analysis of behavior, examines the methods and algorithms used in this area, and identifies problems. Proposals are developed for the effective solution of questions on the intellectual analysis of behavior in the e-government environment. The approach we propose for intellectual analysis of behavior based on textual information consists of 4 levels: 1) primary processing, 2) document description, 3) classification of a set of documents into positive and negative classes, 4) determination of accuracy and completeness characteristics in classification. The use of semantic indicators for intellectual analysis of behavior can help conduct research with greater accuracy and effectively solve behavioral prediction problems (pp.64-72).