Objective To systematically evaluate the effectiveness and safety of fluoxetine in treating premature ejaculation (PE). Methods All randomized controlled trials (RCTs) on fluoxetine treating PE published from July 1996 to May 2012 were collected in the following databases: MEDLINE, EMbase, PubMed, Ovid, The Cochrane Central Register of Controlled Trials, CBM and CKNI. According to the inclusion and exclusion criteria, literature screening, data extraction and quality assessment were conducted independently by two reviewers. Then meta-analysis was performed using RevMan 5.0 software. Results A total of 6 RCTs involving 221 patients were included finally. The results of meta-analysis showed that, as for effectiveness, there was no significant difference in the intravaginal ejaculatory latency time (IELT) between the two groups before the treatment (WMD=–0.21, 95%CI −4.79 to 4.37, P=0.93), but the IELT of the fluoxetine group was obviously longer than that of the control group after the treatment, with a significant difference (WMD=134.54, 95%CI 79.78 to 189.30, Plt;0.000 01). The results of sensitivity analysis indicated that the IELT of the fluoxetine group was longer than that of the control group, with a significant difference (WMD=155.19, 95%CI 130.64 to 179.75, Plt;0.000 01). As for safety, the fluoxetine group was higher in the incidence of adverse reaction than the control group, with a significant difference (OR=5.49, 95%CI 2.43 to 12.38, Plt;0.000 1). Conclusion Current evidence indicates that fluoxetine can improve the symptoms of PE patients, obviously prolong the IELT, and improve the quality of sexual life; and it is tolerable to patients with mild adverse reactions and is suitable for long-term intake. For the limited quantity of the included studies, we herein believe that, to obtain more evidence, it is necessary to further confirm the diagnosis and therapeutic criteria of PE, to design and conduct more multicenter and large scale clinical studies by adopting the internationally recognized indexes, and to perform a long-term follow-up．