Understanding the underlying processes of how communities are structured remains a central question in community ecology. However, the mechanisms of the soil animal community are still unclear, especially for communities on a small scale. To evaluate the relative roles of biotic interactions and environmental and spatial processes in a soil collembolan community, a field experiment was carried out on a small scale(50 m) in the farmland ecosystem of the Sanjiang Plain, Northeast China. In August and October, 2011, we took 100 samples each month in a 50 m × 50 m plot using a spatially delimited sampling design. Variation partitioning was used to quantify the relative contributions of the spatial and environmental variables. A null model was selected to test for the non-randomness pattern of species co-occurrence and body size in assemblages of collembolans and to test whether the pattern observed was the result of environmental or biotic processes that structured the community on a small scale. The results showed that large variance was accounted for by spatial variables(18.99% in August and 21.83% in October, both were significant). There were relatively lower effects of environmental variation(3.56% in August and 1.45% in October, neither was significant), while the soil water content, soil p H and soybean height explained a significant portion of the variance that was observed in the spatial pattern of the collembolan community. Furthermore, the null model revealed more co-occurrence than expected by chance, suggesting that collembolan communities had a non-random co-occurrence pattern in both August and October. Additionally, environmental niche overlap and the body size ratio of co-occurrence showed that interspecific competition was not influential in collembolan community structuring. Considering all of the results together, the contributions of spatial and environmental processes were stronger than biotic interactions in the small-scale structuring of a soil collembolan community. Understanding the underlying processes of so communities are structured remains a central question in community ecology. However, the mechanisms of the soil animal community are still unclear, especially for communities on a small scale. To evaluate the relative roles of biotic interactions and environmental and spatial processes in a soil collembolan community, a field experiment was carried out on a small scale (50 m) in the farmland ecosystem of the Sanjiang Plain, Northeast China. In August and October, 2011, we took 100 samples each month in a 50 m × 50 m plot using a spatially delimited sampling design. Variation partitioning was used to quantify the relative contributions of the spatial and environmental variables. A null model was selected to test for the non-randomness pattern of species co-occurrence and body size in assemblages of collembolans and to test whether the pattern observed the result of environmental or biotic processes that structured the community on a smal l scale. The results showed that large variance was accounted for by spatial variables (18.99% in August and 21.83% in October, both were significant). There were relatively lower effects of environmental variation (3.56% in August and 1.45% in October, neither was significant, while the soil water content, soil p H and soybean height explained a significant portion of the variance that was observed in the spatial pattern of the collembolan community. Furthermore, the null model revealed more co-occurrence than expected by chance , suggesting that collembolan communities had a non-random co-occurrence pattern in both August and October. Additionally, environmental niche overlap and the body size ratio of co-occurrence showed that interspecific competition was not influential in collembolan community structuring. Considering all of the results together, the contributions of spatial and environmental processes were stronger than biotic interactions in the small-scale structuring of a soil collembolan community.