The present study classifies target users of Central Library of Isfahan to identify their behaviors and information needs. The research uses applied and descriptive-survey method with quantitative approach using neural network techniques in data-mining. The research tool was a researcher-made questionnaire including 64 items with the average reliability of Cronbach’s alpha coefficient 0.862. The sample population consisted of 356 faculty members and graduate students of Isfahan University of Technology in the year 1393-94. Using MATLAB, they were clustered based on each of the main components of the study (the purpose and motivation of information-seeking, the amount of using information sources, information-seeking obstacles and skills, and the amount of using information services). Removing each of the research’s sub-components of the main components demonstrated the highest purpose and motivation of information-seeking among the target population, and the least effective were reaching a scientific level and competition with rivals; the information sources mostly needed by library users were translations of Latin books, and their minimum requirements were reports and research schemes; the main barrier to their information-seeking was lack of access to proper Internet and the least effective factor in lack of access to information sources was inconvenient use of the sources; the skill of awareness about the required information and access to information channel are most efficient items in the information seeking skills. The most inefficient item is the ability of users in recognizing the importance of obtained information from their searches with respect to their required information; the most effective services in meeting the information needs of the users were the ability of librarians and information science experts in information retrieval and their knowledge of information services and resources, and the least effective were familiarity with a variety of information services and resources in their professional major and possibility of using the information sources in the form of digital files needless of attending libraries. According to the purpose and results of the present study, it is possible to classify target users of libraries via using neural network technique in data-mining, based on their needs and information-seeking behaviors.