In order to evaluate the pollution characteristics and sources of As and other heavy metals in farmland around Huize lead-zinc mining area in Yunnan Province, 42 soil samples were collected and nine elements in the soils were determined. The absolute principal component scores/multiple linear regression (APCS-MLR) receptor model was used to quantitatively analyze the contribution rate of soil As, and the geostatistical method was used to reveal the different sources of As and its spatial distribution. The results showed that:1) the average content of As in the study area was 73.3 mg/kg, which was 1.8 times higher than the risk screening values for soil contamination of agricultural land, 2) according to APCS-MLR receptor model, the contribution rates of As pollution sources were calculated as source I (58.6%), source II (14.7%) and other sources (26.7%), 3) based on PCA and the spatial distribution patterns of the source contributions, source I may originate from the mining related activities, source II may come from coal combustion and fertilization, other sources may be natural sources such as geological background. This study indicated that combining the APCS-MLR receptor model and geostatistics was an effective method for apportioning soil heavy metal pollution sources and contribition.