The researches on the plant population spatial pattern were mostly based on 0-D plant point (0-D IND) or 0-D plant count, and only a few was based on the 2-D projective cover (2-D PC) and 3-D aboveground biomass (3-D AGB reflected by canopy volume). Until now, the plant population spatial distribution patterns incarnated by these parameters were still unclear. Taking the widely distributed small semi-shrubs Seriphidium terrae-albae and Artemisia songarica in Gurbantunggut Desert of Northwest China as test objects, this paper studied the IND, PC, and AGB of each individual at two sampling plots. Through six-scale division of plot coordinate system with GIS, and by using aggregation analysis, coefficient of variation (CV) , and a scaling exponent between the CV and six scales, the characteristics of the population spatial distribution patterns with the above mentioned parameters were comparatively analyzed. At all scales, the IND (except for the S. terrae-albae population at 0.5 m scale) and the AGB of the two shrubs all presented a clumped distribution, and the aggregation intensity increased' with increasing scale. However, the PC had a uniform distribution (except for the A. songarica population at 5 and 8 m scales). With increasing scale, the CV values of the two shrubs decreased. The absolute value of scaling exponent (k value) of the IND was higher than those of the PC and AGB, and there was no significant difference in the k values between the PC and AGB, indicating that the scale variation scope of the struc- tural complexity of the IND was larger than that of the PC and AGB. The k value of each parameter for S. terrae-albae was higher than that for A. songarica, which could be related to the populations' interspecific relationship and plant size. In sum, the IND and AGB had similar spatial patterns, while the PC and AGB had almost same spatial pattern complexity and scale change characteristics.