The current study examined the effectiveness of brief cognitive behavior therapy (CBT) for severe mood disorders in an acute naturalistic setting. The sample included 951 individuals with either major depressive disorder (n = 857) or bipolar disorder with depressed mood (n = 94). Participants completed a battery of self-report measures assessing depression, overall well-being, and a range of secondary outcomes both before and after treatment. We found significant reductions in depressive symptoms, worry, self-harm, emotional lability, and substance abuse, as well as significant improvements in well-being and interpersonal relationships, post-treatment. Comparable to outpatient studies, 30% of the sample evidenced recovery from depression. Comparison of findings to benchmark studies indicated that, although the current sample started treatment with severe depressive symptoms and were in treatment for average of only 10 days, the overall magnitude of symptom improvement was similar to that of randomized controlled trials. Limitations of the study include a lack of control group, a limitation of most naturalistic studies. These findings indicate that interventions developed in controlled research settings on the efficacy of CBT can be transported to naturalistic, "real world" settings, and that brief CBT delivered in a partial hospital program is effective for many patients with severe depressive symptoms.