Abstract: In recent years, particulate matter (PM_2.5) concentrations in China have decreased significantly; however, there are notable temporal and regional differences in the changes of its components. Current research on PM_2.5 composition, primarily based on numerical models and online observations, mostly focuses on single sites and representative periods, lacking high spatiotemporal resolution analyses of regional and interannual variations. This study employs a multivariate tracer model, combined with gridded air quality monitoring data, to construct a dataset of primary PM_2.5 (PPM) and secondary PM_2.5 (SPM) concentrations for 367 cities in China from 2015 to 2023. The results show that PPM and SPM concentrations decreased from 15.85 μg m⁻³ and 25.86 μg m⁻³ in 2015 to 10.28 μg m⁻³ and 15.42 μg m⁻³ in 2023, representing reductions of 35.02% and 40.37%, respectively, over 9 years in China. The interannual trends of PPM and SPM from 2015 to 2023 showed different characteristics before and after 2021, with a rapid decline from 2015 to 2021 (−0.91 μg m⁻³ yr⁻¹ and −1.70 μg m⁻³ yr⁻¹, respectively). After 2021, the decline in PPM slowed (−0.14 μg m⁻³ yr⁻¹), whereas SPM showed a slight increase (+0.10 μg m⁻³ yr⁻¹). The overall trend of SPM% was similar to that of SPM, decreasing at a rate of −0.38% yr⁻¹ before 2021 and increasing at a rate of +0.38% yr⁻¹ after 2021. Regionally, the highest concentrations of PPM and SPM were found in North and Central China, corresponding to stronger anthropogenic emissions. However, SPM% was higher in South and Southwest China (66.67%–68.66%), likely due to the stronger atmospheric oxidation capacity in these regions. The trends in PPM and SPM concentrations across most regions were consistent with the national trend, except in the Southwest region, where both SPM and SPM% continued to decline significantly after 2021. This study offers scientific evidence and data support for future particulate matter pollution control in China.