We tested the hypothesis that exercise would improve subjective sleep quality and activity in depressed elders. A 10-week randomized controlled trial was utilized. Participants consisted of a volunteer sample, aged > 60 with a diagnosis of major or minor depression or dysthymia. A total of 32 subjects aged 60-84 years with a mean age of 71.3 +/- 1.2 years was used. Intervention consisted of a supervised weight-training program three times a week or an attention-control group. Main outcome measures were Pittsburgh Subjective Sleep Quality Index (PSQI), Likert Scale of Subjective Sleep Quality and Quantity. Paffenbarger Activity Index. Geriatric Depression Scale (GDS). Beck Depression Inventory (BDI), Hamilton Rating Scale of Depression (HRSD), and the Medical Outcomes Survey Short Form 36 (SF-36). Results showed that exercise significantly improved all subjective sleep-quality and depression measures. Depression measures were reduced by approximately twice that of controls. Habitual activity was not significantly increased by exercise. Quality of life subscales significantly improved. In a forward stepwise multiple regression, percent improvement in GDS and percent increase in strength remained significant predictors of the improvement in total PSQI score (r = 0.71, p = 0.0002). In conclusion, weight lifting exercise was effective in improving subjective sleep quality, depression, strength, and quality of life without significantly changing habitual activity.