We examined a multilevel system for forecasting industry performance. We are looking at an industry that is considered a large-scale system with many connections. Modern means of collecting and processing large data arrays make it possible to solve the problem of information support of models. We use open data from Russian organizations for modeling and forecasting their activities. We consider the problem of forecasting in the inertial version, after which scenarios (measures) for improving performance are considered. An example of forecasting the financial and economic indicators of the industry is considered using the example of air transport. Calculation results show an acceptable level of accuracy.