Mannitol-yolk-polymyxin B agar (MYPA) was modified by supplementation with trimethoprim. The ability of the supplemented medium to select for and recover Bacillus cereus from pure cultures and food samples with high background microflora was compared with MYPA. For evaluation of the modified MYPA (mMYPA) in food samples with high background microflora, B. cereus was experimentally spiked into red pepper powder, fermented soybean paste, vegetable salad, and radish sprouts, and then it was recovered on MYPA and mMYPA for comparison. In all food samples, there was no difference in recoverability (P > 0.05) between mMYPA (red pepper powder, 3.34 ± 0.24 log CFU/g; fermented soybean paste, 3.52 ± 0.47 log CFU/g; vegetable salad, 3.51 ± 0.23 log CFU/g; radish sprouts, 3.32 ± 0.40 log CFU/g) and MYPA (red pepper powder, 3.18 ± 0.20 log CFU/g; fermented soybean paste, 3.33 ± 0.43 log CFU/g; vegetable salad, 3.36 ± 0.19 log CFU/g; radish sprouts, 3.33 ± 0.31 log CFU/g). However, mMYPA exhibited better selectivity than MYPA, because additional trimethoprim made the differentiation of suspected colonies easier by inhibiting competing flora. The addition of trimethoprim to conventional media could be a useful option to improve selectivity in foods with high background microflora.