The Chinese government has implemented a clean heating policy to alleviate the severe air pollution and high energy consumption associated with winter heating in northern China. However, the promotion of clean heating technology paths (CHTPs) is limited by the problems of rising energy prices and insufficient supply capacity. To quantitatively study the promotion scope of each CHTP, this paper adopts the K-means clustering method in machine learning to analyze the clean heating methods suitable for the characteristics of each region in northern China based on 37 pilot cities. The results show that the 10 CHTPs based on coal, electricity and natural gas as energy, the most widely applied is Air-source heat pump technology, accounting for 86.7 % in cities. The least number of cities suitable for application are water-source heat pumps and sewage-source heat pumps due to economic constraints, with a proportion of 44 %. The 3 kinds of CHTPs applying solar, biomass and geothermal energy, accounted for more than 60 %, among them, ground-source heat pumps are recommended to promote the smallest scope, concentrated in Hebei, Henan, Shandong, and Heilongjiang provinces. This study is of great significance in guiding the selection of CHTPs on a large spatial scale.