BACKGROUND: We assessed the association between dietary diversity and CD4 count, moderate anemia, and mortality among 876 antiretroviral therapy-naive people living with HIV/AIDS infection (PLHIV) in Uganda. METHODS: Participants were interviewed and followed for an average of 21.6 months. Dietary diversity was measured using the Individual Dietary Diversity Score (IDDS) (range, 0-12) and summarized into an overall measure and disaggregated into nutrient-rich food groups (range, 0-7), cereals, roots, and tubers (range, 0\x{2013} 2); and oils, fats, sugars, and condiments (range, 0\x{2013} 3). We determined the cross-sectional associations between dietary diversity and (1) immunosuppression (CD4 count ≤ 350 cells/μL) and (2) moderate anemia (hemoglobin < 10 g/dL) at baseline with logistic regression. We assessed the association between IDDS and mortality using Cox proportional hazards regression. RESULTS: The mean IDDS score was 6.3 (SD 1.7) food groups per day, with a mean of 2.7 (SD 1.1) nutrient-rich food groups per day. Each additional nutrient-rich food group consumed was associated with a 16% reduction in the likelihood of having a CD4 count ≤350 cells/μL [adjusted odds ratio, 0.84; 95% confidence interval (CI): 0.72 to 0.97] at baseline. Among those with >350 CD4 cells per microliter, but not those with CD4 count ≤350 cells per microliter, consumption of nutrient-rich food groups was associated with a lower odds of moderate anemia (adjusted odds ratio, 0.57; 95% CI: 0.34 to 0.96). During follow-up, 48 participants (5.6%) died (mortality rate of 3.1 per 100 person-years). IDDS was inversely associated with mortality [adjusted hazard ratio, 0.76; 95% CI: 0.63 to 0.91]. CONCLUSION: These results suggest that diet quality is an important determinant of HIV disease severity and mortality in antiretroviral therapy-naive PLHIV.