In order to improve the safety and intelligence level of forklift truck，a forklift stability warning method based on multi-core parallel digital twin is proposed. First，the open-source high-precision inertial navigation sensor is used to evaluate the driver's operating habits and road smoothness，and then drive the digital twin operation. Secondly，the optimization algorithm is used to calculate the limit operating parameters that meet the stability conditions，such as the highest load and the fastest driving speed. Then，two acceleration strategies are proposed，which use the open source C language compiled solver and multi-core Central Processing Unit (CPU) in parallel，to shorten the optimization iteration time. Finally，taking a forklift truck as the test object，the accuracy of the measurement of the flight control system is verified，and the real-time performance of the digital twin stability early warning method is analyzed. The results show that the flight control system has high monitoring accuracy，the theoretical calculated values are basically consistent with the collected values，and the longitude and latitude deviation of global positioning system (GPS) is about 1 m. The digital twin program has a faster prediction speed. The digital twin compiled by the open source C language is 3 times faster than the Simulation X3.8 platform，and the parallel digital twin method is 20 times more efficient than the simulation platform.