Background: The purpose of this study was to compare operative mortality and long-term outcome of patients undergoing tricuspid valve replacement versus tricuspid valve repair. Methods: From February 1986 to July 2006, 315 patients underwent tricuspid valve surgery including 93 replacements (72 biologic, 21 mechanical) and 222 repairs. To control for selection bias and varying comorbidities, a matched cohort of patients undergoing repair versus replacement was selected using propensity score analysis (68 patients in each group). Results: In the propensity-matched cohorts, operative mortality was similar for tricuspid valve replacement (13% ± 4%) and repair (18% ± 5%; p = 0.64). Intensive care unit length of stay was similar between cohorts (replacement, 4 days; repair, 3 days; p = 0.45), but the replacements had a significantly longer hospital lengths of stay (9 days versus 6 days; p = 0.01). In the replacement cohort, survival was 85% at 1 year, 79% at 5 years, and 49% at 10 years. In the repair cohort, survival rates were similar with 80% at 1 year, 72% at 5 years, and 66% at 10 years (p = 0.66 versus replacement). Conclusions: Surgical treatment of tricuspid valve disease, regardless of the operative approach, is associated with significant early and late mortality. However, there is no difference favoring tricuspid valve repair over replacement. Thus, we should not hesitate to consider tricuspid valve replacement for patients in whom we believe there is a reasonable chance for recurrence of regurgitation after repair.