Design/methodology/approach The assessment of the rationality of motor oil change intervals for automobiles operating under urban conditions was systematically conducted through an examination of physicochemical properties and tribological performance using a four-ball tester. The results demonstrated that the physicochemical parameters of the used oil, including total acid number, oxidation onset temperature (OOT) and kinematic viscosity, had not yet reached the threshold levels specified by standards for replacement. Furthermore, there was no significant decline observed in antifriction and wear resistance properties. Findings National Key Research and Development Program of China (Grant No. 52405227, No. 52305210). Originality/value Based on an established relationship model between service mileage of automobiles and OOT, back propagation neural network models demonstrate high accuracy in fitting data with better generalizability; expected model values exhibit a high degree overlap with test values. This study provides fundamental data and technical support for scientifically assessing appropriate intervals for changing automobile oils. Peer review The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-12-2024-0467/