Oil plays an important role in ensuring national energy security. It is of great significance for China’s energy policy-making to predict the future oil demand scientifically. Combined with the grey correlation analysis and combination model of ARIMA-BP neural network, data on six driving variables of China’s GDP, international oil price, energy consumption per unit GDP, industrial structure, urbanization rate and oil production in 1999-2018 were selected to predict China’s oil demand under different scenarios in 2020-2030. The results show that the selected driving variables are highly correlated with China’s oil demand, and the combined model is more accurate than the single model. From 2020 to 2030, China’s oil demand and import will continue to increase, but the growth rate will gradually slow down. Under the three scenarios, the predicted average oil demand in 2020, 2025 and 2030 will be 675.7703 million tons, 732.2725 million tons and 760.8155 million tons, respectively; the predicted average external dependence in 2020-2030 will be 76.19%, far exceeding the international warning line of 50%. The contradiction between China’s oil supply and demand will be more acute, and accordingly, some policy suggestions were put forward.