PM2.5 is related to urban air quality, while human health and landscape pattern of land use is one of the key factors affecting PM2.5. As a spatial carrier of PM2.5, landscape pattern is inevitably affecting local and regional PM2.5 concentration. It is of great significance to explore the influence mechanism between landscape pattern and PM2.5 for environmental protection and air pollution control. Based on PM2.5 monitoring station data and auxiliary data, the geo-weighted regression model (GWR) was used to simulate conditions on the PM2.5 concentration in the Guanzhong Plain Urban Agglomeration from 2015 to 2019. Then the landscape sprawl index(CONTAG) was selected by the principal component analysis, and introduced into the GWR model to analyze the influence of landscape pattern index on the PM2.5 concentration simulation. The accuracy of the PM2.5 concentration modelling was validated with the observed data from air quality monitoring stations and third-party air pollution dataset (CHAP) respectively.The results show that CONTAG added makes the GWR model more suitable for the characteristics of the field-observed data itself with better validation accuracy and stability, and improves the goodness of fit by 4.65%. It was indicated that the GWR model with landscape index incorporated can simulate PM2.5 concentration in Guanzhong Plain Urban Agglomeration with high spatial resolution and accuracy. The results not only conduce to understand the influencing factors and spatial heterogeneity of PM2.5 especially the comprehensive impact of landscape pattern on PM2.5 in Guanzhong Plain Urban Agglomerations, but also sever the purpose of providing scientific support for air pollution prevention at the landscape scale in urban agglomerations.

landscape pattern index; PM2.5; geo-weighted regression; Guanzhong Plain Urban Agglomeration