Using Geographic Information System (GIS) technology and geostatistics (GS) methods followed by simulation, we analyzed the spatial structure and distribution of Bemisia tabaci (Gennadius) in Xinjiang. We also conducted a parallel analysis using traditional statistics. Results from the two analyses revealed that the spatial distribution patterns of B. tabaci (Gennadius) on cotton were aggregated; that the aggregation degree fluctuated over time; and that the proportion of spatial variation had a random component. Using frequency distribution goodness of fit tests, the negative binomial distribution with its parameter set at zero frequency had the closest fit for predicting this species′ spatial distribution in Xinjiang. Results from the GS analysis of the variation in distribution areas and of adult diffusion patterns suggested that the adults of B. tabaci (Gennadius) were spatially correlated in cotton fields. The degree of randomness ranged from 19.22% to 49.99%; the distance range of spatial dependence varied from 32 min to 6372 min July and then fluctuated between a distance of 2000m and 3000m. B. tabaci (Gennadius) populations initially invade cotton fields downwind from their over-wintering areas. Thereafter, the distribution was largely affected by wind direction. It is a classic case of insect distribution caused by winds. Vertical distribution on cotton plants of B. tabaci (Gennadius) was influenced strongly by the effects of the warm temperate zone and of the continental arid desert climate in Turpan. The population initially damages the top leaves, and then proceeds to feed on leaves in the middle and bottom part of cotton plants in the field of downwind, whereas the damage will start with the middle and bottom part of cotton plants in the field of upwind. Once a population is established, its success is based on its adaptability of B. tabaci (Gennadius) to the characteristic climate of the Turpan region.