Poplar Populus davidiana (Salicaceae) is one of pioneer species and a major timber species in northern China. Its sex ratio, spatial distribution and the relationship with environmental factors were studied over three consecutive years in a temperate secondary forest in Changbai Mountain, northeastern China. All trees with DBH≥1cm within a 5.2 hm² (200m×260m) plot were mapped, and all reproductive poplar trees were sexed. In total, 433 poplar trees with dbh above 5 cm were recorded, including 222 males, 183 females, 26 dead trees, and 2 trees with unclear sex idendity. Almost all trees larger than 10 cm were flowered. The size distribution and the relative size variation were not significantly different between sexes. There were no differences in the threshold sizes for flowering between males and females. Although males outnumber than females in many classes, sex ratio of reproductive trees did not differ significantly from 1 ∶ 1. Nearest neighbor analyses indicated that although trees tended to have an individual of the same sex as their neighbor, there was no significant spatial segregation or clumping between individuals of either sex. A univariate O-ring statistical analysis showed that under the complete spatial randomness null model, males and females occurred as clumped distributions at small scales, while under the heterogeneous Poisson process null model, both males and females were distributed randomly at 0 100 m scales. We conclude that the poplar spatial pattern was similar to one of the heterogeneous Poisson process. Clumping of the same sex trees at small scales was partially related to cloning. The bivariate O-ring statistics indicated that under the antecedent condition, big trees attracted small trees of the same sex within 18 m spatial scales, big trees independent from small trees of the opposite sexes; under independence null model, males and females independent from each other. Apparently, there is no difference in the spatial occupation or use of the environment between trees of either sex, although the differential use of the environment has been proposed as one of the evolutionary advantages of dioecy; To use the space as an explanatory variable, we applied the Principal Coordinates of Neibourhood Matrices method. Variation partitioning with regression analyses was used to discover which variables better explained variation in the spatial pattern of males and females. Stand density was the only selected environmental factor by 999 permutation forward selection to perform variation partitioning both for males and females. The variation partitioning of the spatial pattern indicated that stand density showed almost no impact on females, litter effect (8.19%)on males, but spatial variables showed stronger effect (high to 48%) on females compared 13.65% on males. The combination of environmental factor and spatial variables explained 27% of males' spatial distribution, much less compared to 52% of females'. The study results suggest that spatial patterns of poplar may result from some processes not analysed in the current study, as seed dispersal, interspecies competition, and/or disturbance.