Little is known about the efficacy of sampling methods for monitoring bee communities in agricultural landscapes in sub-Saharan Africa, in general, and in Uganda, in particular. To provide baseline information on the effectiveness of different sampling methods, 26 sites with varying landscape characteristics were sampled in 2006 in agricultural landscapes in Uganda. Bees were sampled using line transect count, coloured pan trap and hand net methods. In total, 80,883 bee individuals were collected. Totals of 59, 314 and 559 bee species were recorded in transect counts, pan traps and hand nets, respectively. Thus hand nets captured the most species overall. There were few species that overlapped across the three sampling methods. Wild bees were significantly more abundant in yellow pan traps than in blue or white ones. In contrast, bee species richness was significantly higher in blue pan traps than in white or yellow pan traps. Overall, pan-trapping was found to be a complementary method to hand-netting for monitoring bee communities in Uganda.