The article describes a new method for making management decisions on heat supply in district heating systems, based on solving a sequence of recurrence relations of fi rst-order differential equations, enabling to synthesize daily schedules of heat supply in such systems. Using fi rst order differential equations, we implement real-time daily heat supply scheduling, predict the time-temperature dependence for heating water in the supply line, and we form a decision on the thermal energy delivery on the basis of this information. The effectiveness of our method is confi rmed by numerical modeling and comparative analysis of daily heat supply scheduling with the help of advanced intelligent decision making tools. For comparative analysis, we considered daily scheduling using a nonlinear regression model, a generalized regression neural network, a radial basis neural network, and a linear neural network. The effectiveness of our method was estimated on the basis of MAPE (mean absolute percentage error) and Accuracy coeffi cients. The model was recognized as most effective for which the MAPE value was maximum, and the Accuracy value tended to one hundred percent. Experimental studies showed that our proposed model has an advantage over the regression model by 1.68 times and over the neural models by more than 10.2 times when modeling for a hundred heating network sections. Thus, the main purpose of our study was to increase the accuracy of the method of making a managerial heat supply decision based on the experimental verifi cation of a mathematical model of thermal inertia of a branched district heating system.