Abstract: In order to investigate the causes of atmospheric particulate matter pollution in spring in a dense nonferrous smelting area in Yunnan Province, 9 sampling points were set up in Mengzi City, Gejiu City and Kaiyuan City (Ge-Kai-Meng Region), the main dense nonferrous smelting area of Yunnan Province, from March 20^th to 29^th, 2017, to simultaneously collect PM_2.5 and PM₁₀ ambient samples. Nineteen chemical components, such as inorganic elements, water-soluble ions, carbon compositions, were detected. The physical and chemical characteristics of the particles were analyzed. The sources of atmospheric particles in Ge-Kai-Meng Region were described based on enrichment factor (EF), backward trajectory model, and principal component analysis and multivariate linear regression (PCA-MLR). The results showed that: ① During the sampling period, the mass concentration mean values of PM_2.5 and PM₁₀ were (50.79±20.67) μg/m³ and (73.47±33.86) μg/m³, and the mass concentration of PM_2.5 exceeded standard Ⅱ (75 μg/m³) on March 24^th. ② The EF values of Cu, Zn, Ca, Pb, As, Cd, etc. were 10—4000 in PM_2.5 and PM₁₀, which were greatly affected by anthropogenic emission sources. ③ The result of backward trajectory model showed that the air masses arriving in Ge-Kai-Meng Region in those polluted days mainly passed through the central part of Myanmar with the dense biomass burning points, which indicated that the emission of biomass burning from Southeast Asia in spring significantly affected the air quality of Ge-Kai-Meng Region. ④ The result of PCA-MLR model showed that biomass combustion and nonferrous smelting activities had important contributions to atmospheric particulate matter during the sampling period in Ge-Kai-Meng Region, among which, the contribution rate of biomass combustion was over 30%, and the composite contribution rate of nonferrous smelting activities was over 45%.