The necessity to determine the technical condition of automotive internal combustion engines arises when their power, economic and environmental indicators deteriorate, as well as before performing the next major repair. Currently, disassembly and non-disassembly methods are used for this purpose. Disassembly methods, although they are the most reliable, are more labor-intensive and require the removal of the engine from the car, the availability of space and equipment for its disassembly. And non-disassembly methods have a low labor intensity, they practically do not require removal of the engine, while providing different reliability of the diagnosis depending on the method used. A method of non-disassembly diagnostics of automotive engines based on the pulse functions of their systems is proposed. As diagnostic parameters here are offered the pulse function of changes in the fuel and air supply of the engine at known values of compression in the cylinders. At the same time, the experience of operating automotive engines shows that the pulse function is a function of the current technical condition of the engine. By using the generalized Bayes formula for determining conditional probabilities, it is possible to determine the probability of failure of engine systems and the current technical condition. The undoubted advantage of the proposed method is its trainability, i.e. with an increase in the statistical data of the diagnosed engines, the reliability of the established diagnosis increases. The offered method will be useful for engineers and technicians involved in the maintenance of internal combustion engines for practical implementation.