Objective To evaluate the efficacy and safety of topical retinoids in the treatment of acne vulgaris compared with placebo, antibiotics, benzoyl peroxide and sulfur preparation. Methods According to the Cochrane reviewer′s handbook, randomized controlled clinical trials were selected for the systematic review. Results Up to 2002, 15 clinical trials (2,439 patients) that met the inclusion criteria were selected. There were four clinical trials which showed that topical retinoids were more effective than that of placebo (RR=1.87, and 95% CI: 1.13 ~ 3.11),especially for noninflammatory lesions (RR=12.70,and 95% CI : 4.09 ~ 39.40). There were 3 clinical trials which showed that topical retinoids had better efficacy than that of sulfur preparations (RR=1.75, and 95% CI: 1.42 ~ 2.16). For 7 clinical trials of retinoids compared with benzoyl peroxide, and 3 clinical trials of retinoids compared with antibiotics, no conclusion could be drawn. All the clinical trials showed that there were local side effects, including erythema, and scaling etc in the patients using topical retinoids, but no systematic side effects were observed, however, pregnant women had to be very cautious. Conclusions Topical retinoids are effective for acne vulgaris, and has better efficacy than sulfur preparation does, but there is not enough evidence to clarify that the efficacy of topical retinoids is better than that of benzoyl peroxide and antibiotics.