Taking the price data of second-hand houses and the POI data of urban service facilities in Chengdu from 2013 to 2021 as an example, the spatio-temporal relationship between house prices and urban service facilities under the spatial grid scale of 1 km×1 km is explored. The results show that spatially, the distribution of urban service facilities and living spaces was consistent, and the overall "meter" radial features along the traffic trunk line are the same.The correlation between the distribution of urban service facilities and housing prices was significant, but it was gradually weakening, especially the decline in the correlation coefficient of for-profit facilities was more obvious, and the differentiation of housing prices within the region was deepening. The geographical detector model could effectively identify the differences in the impact of different variables on housing prices, with the strongest impact on population density, building year, and road network density, and the interaction between primary education facilities and other types of interaction has higher interpretation of housing prices.