The work is studying the distributed system of natural environment monitoring points. The purpose of the study is to solve two main problems of optimization: a) optimization of risks of emergency events; b) optimal cessation of environmental monitoring, which allows to reduce the costs of observation. Monitoring of the environment contaminated with petroleum products is relevant. Traditionally used methods and monitoring itself are expensive and technological complex mechanisms, often requiring satellite data. When studying environmental pollution, it is important to consider situational risks and stochastic irreversible changes in nature. For this purpose, it is necessary to apply the methods and methodology of the system approach, which are used in the article for analysis of complex data structures and entropy of the system. For the informativeness of less long monitoring, it's represented by four main subsystems (capture, operation, selection and decision-making) and evaluated using an information-entropy approach The work will solve the practical task of monitoring the contaminated environment.