Under the background of the reform of energy development strategy and low oil prices, the sustainable development of oilfields faces multiple challenges of new energy squeeze, energy conservation and emission reduction, and low-cost development. Since oilfield development is a fairly complex process, only a more accurate prediction of oilfield output can better plan oilfield exploitation, reduce production investment, improve productivity, and promote the sustained and steady development of the oilfield economy. Based on summarizing the common oilfield production forecasting models: Weng’s model, decreasing yield model, time series method, machine learning method, etc., this paper analyzes and compares its similarities and differences, and proposes that the method based on the combined forecasting model is the development trend of oilfield production forecasting under the idea of low-carbon development in the future.