The article presents a new method for solving the scientific and practical problem of automated calculation of the weight of general-purpose bridge cranes. Numerical data on the weight of existing cranes are given in tables and structured depending on the load capacity and span. Mathematical statistics’ hypotheses were used, making it possible to distribute the results according to the normal law under the same accuracy of the obtained data. Based on these assumptions, the least squares method was applied, which made it possible to construct a function of two variables that determines the dependence of the crane weight on the span and load capacity, combining these parameters. A formula was obtained that makes it possible to calculate the weight of cranes programmatically. Based on statistical data, a neural network was built, which, similarly to traditional statistical methods, finds the weight of bridge cranes. The quality of the obtained result was assessed using traditional statistical methods and using a neural network. The work of the statistical model and neural network was studied beyond the data definition area. Functional dependencies for determining the weight of overhead cranes in the range of up to 50 tons are determined based on the regression equation and using a neural network. It is established that forecasting is unacceptable outside the range, but the neural network copes better with the task of extrapolation than traditional methods. The article substantiates the advantages of the proposed method.