Knowing the dispersal of released insects and estimating the size of field populations are keys to the successful implementation of area-wide management (AWM) programmes based on the sterile insect technique (SIT), as they determine the release strategy of sterile males. Mark–release–recapture (MRR) is a common method used to estimate field populations and spatiotemporal dynamics. However, the extent to which the pest is attracted to lures is often difficult to identify, thereby biasing extrapolation to movement patterns and population size. We performed MRR experiments on the Oriental fruit fly, Bactrocera dorsalis (Hendel) (Diptera: Tephritidae), in a fruit-growing area in Senegal. Methyl eugenol and protein baits were used to trap males and females, respectively. After studying the suitability of marking B. dorsalis with fluorescent pigments at the laboratory, two releases of marked sterile flies were organized in the centre of an organic mango orchard, first in the late mango fruiting stage and second in the fruit setting stage. Traps were placed symmetrically along a 250 and 500 m radius from the release point. A very small fraction of the released individuals was recaptured: 0.45% in the first release and 3.08% in the second. Trapping of both sterile and wild flies was completely anisotropic. Sterile flies were mostly trapped at a relatively short distance (250 m) from the release point, in the first two days. Male trapping using methyl eugenol was highly effective, whereas the response of females to food bait traps was low. The direction of the wind was the main driver of recapture, with flies heading upwind. The results underline the importance of taking the odour plume around the traps into account when estimating populations, and the heterogeneous spread of the wild population in the landscape for the set-up of the release strategy of sterile insects for SIT-based AWM.