Four mark-release-recapture experiments were conducted from May to September 2000, to construct a 3D dispersal model for aster leafhopper (Macrosteles quadrilineatus Forbes). In the laboratory, flight mills and an optical sensor were used to measure the effects of fluorescent dust and rabbit protein marking on flight activity of aster leafhopper. No significant differences in proportion of leafhoppers flying, distance flown, average flight speed, or wing-beat frequencies were observed among marked leafhoppers versus unmarked controls. Leafhoppers were sampled in a grid pattern around a central release point of marked leafhoppers to estimate the patterns of leafhopper abundance and the distribution of dispersal distances. Geostatistical analysis of numbers of aster leafhopper adults captured by vacuum sampling in the grid pattern around a central release point was used to examine differences in dispersal pattern among crops. The spatial correlation range was ≈200 m for lettuce but only ≈35 m for endive. These differences in spatial pattern suggest leafhoppers disperse more slowly from lettuce plants, preserving aggregations over longer distances. The proportion of leafhoppers recaptured at various distances from a release point was modeled using a normal distribution for dispersal perpendicular to the wind and a Gumbel distribution for dispersal parallel with the wind. Goodness-of-fit tests indicated that both distributions fit field observations well. For all recaptured leafhoppers, the average distance from the release point was 53.6 m and the average distance for dispersal perpendicular to the wind was 17.6 m. The average distances moved parallel to the wind of those leafhoppers dispersing upwind and downwind were 12.1 and 43.9 m, respectively. The dispersal model will be used in spatially explicit simulation of aster yellows epidemiology.