Abstract: The article is devoted to the question whether it is possible to execute modeling and computer-aided evaluation of students' competency levels based on the analysis of their decisions. The emphasis is made on the method of constructing the structure of complicating of routine tasks and operational space of potential solutions. Based on the preset operational space, the authors of the article suggest to create expert agents which solutions will serve as the basis for further analysis of rationality and completeness of students' decisions. As a result of the students' decision anaylsis, the authors develop their competency map. The level of competency is defined by the difficulty of routine tasks soloved by students as part of the given knowledge module. The research involves modeling particular components of the decision making process and interaction methods when making a decision about the level of students' competency. The main conclusions of the research are the following. Firstly, the aforesaid model of the decision operational space, on the one hand, limits a possible set of applicable basic operations and enables formulation and logging of users' decisions by the testing system and, on the other hand, gives freedom to students in the process of constructing potential decisions. Secondly, the array of routine task complication constitutes a good basis for constructing the competency map of a student in a form of an overlay model, allows to model and evaluate the level of competency in accorance with the level of difficulty of tasks being solved, and enables to arrange for adaptive testing that takes into account individual features of students. Thirdly, the data structures introduced in the operational space of decision making can be adjusted in accordance with the incoming information in order to be adequate. The models reviewed create a good basis for implementing multiagent and neural network paradigms for programming intelligence systems of testing.