A simulation study was conducted to examine the effect of pattern of herbicide use on development of resistance to two herbicides with different modes of action in finite weed populations. The effects of the size of the treatment area (analogous to initial weed population), germination fraction and degree of self-pollination in the weed were investigated. The results indicate that the probability of developing resistance to one or both herbicides decreases as the size of the area/initial population decreases. For treatment areas of 100 ha or less with an initial weed seedbank of 100 seeds m−2 and initial frequencies of the resistance genes of 10−6, development of resistance to both herbicides (double-resistance) is uncommon within 50 years for all types of weeds if both herbicides are used in all years (used in combination). If herbicides are used in alternate years (rotated) double-resistance almost always occurs in 100 ha areas but is uncommon in areas of 1 ha or less. The results suggest that adoption of practices that limit movement of weeds in conjunction with using herbicides in combination rather than in rotation can substantially delay development of herbicide resistance.