The mechanisms that drive microbial turnover in time and space have received considerable attention but remain unclear, especially for situations with anthropogenic perturbation. To understand the impact of long-term oil contamination on microbial spatial turnover, 100 soil samples were taken from five oil exploration fields located in different geographic regions across China. The microbial functional diversity was analyzed with a high-throughput functional gene array, GeoChip. Our results indicated that soil microbial α-diversity (richness and Shannon diversity index) decreased significantly with contamination. All contaminated and uncontaminated samples exhibited significant spatial autocorrelation between microbial community similarity and spatial distance, as described by a distance-decay relationship (DDR). However, long-term oil exposure flattened the slopes of the DDRs of all of the functional genes and each functional group involved in C/N/P/S cycling, particularly of those involved in contaminant degradation. The relative importance of deterministic and stochastic processes in microbial assembly was determined. The decrease in microbial spatial turnover with long-term oil contamination was coupled with an increase in the proportion of deterministic processes that structured microbial assembly based on null model analysis. The results indicated long-term oil contamination significantly affects soil microbial community spatial structure by acting as an environmental filter to decrease the regional differences distinguishing soil microbial communities.