Statistical characteristic analysis of pollutants in contaminated sites can help identify the origin, generation, and spatial variation of different pollutants, and can reduce the uncertainty of site survey data. Taking a large and abandoned contaminated coking site of China as the object, 114 surface (0-50 cm) soil samples were collected, with the statistical and spatial characteristics of 16 priority PAHs (sigmaPAHs) analyzed. The descriptive statistical analysis indicated that the sigmaPAH levels varied significantly, and the data were severely skewed. The correlation matrix (CM) and principal component analysis (PCA) showed that the extracted first two principal components (PCs) could effectively represent the whole site pollution data. Four pollutants, i. e., Baa, Bbf&Bkf, Bap, and Inp, were selected for trend analysis and spatial local variance analysis. In the east-west and north-south directions of the site, the pollution showed a low-high-low trend. The variation coefficient of pollution for the site was higher in the central, northwest, and southwest regions, while lower in the other regions.