The purpose of this study was to compare the efficacy and safety of olanzapine (OLZ) monotherapy and an olanzapine/fluoxetine combination (OFC) with placebo (PLA) for unipolar major depression with psychotic features. Under a single protocol, two 8-week, double-blind trials were conducted at 27 sites. Patients (n = 124 trial 1, n = 125 trial 2) were randomized to 1 of 3 treatment groups: OLZ (5 to 20 mg/d), PLA, or OFC (olanzapine 5 to 20 mg/d + fluoxetine 20 to 80 mg/d). The primary outcome measure was the 24-item Hamilton Depression Rating Scale total score. For trial 1, endpoint improvement for the OLZ group (−14.9) was not significantly different from the PLA or OFC groups. The OFC group had significantly greater endpoint improvement (−20.9) than the PLA group (−10.4, P = 0.001); this significant difference was present within 7 days of therapy and maintained at every subsequent visit. The OFC group also had significantly higher response rate (63.6%) than the PLA (28.0%, P = 0.004) or OLZ (34.9%, P = 0.027) groups. For trial 2, there were no significant differences among treatment groups on the 24-item Hamilton Depression Rating Scale total scores or response rates. The combination exhibited a comparable safety profile with OLZ monotherapy and no significant increases in extrapyramidal symptoms compared with placebo. Patients with major depression with psychotic features treated with OLZ monotherapy did not demonstrate significant depressive symptom improvement compared with placebo in either trial; however, an olanzapine/fluoxetine combination was associated with significant improvement compared with placebo in one trial and was well tolerated.