Background and aims. Undernutrition coexists with tuberculosis and is associated with adverse treatment outcomes. Nutrition packages have been incorporated into tuberculosis patient care in some regions but there are little data on its effectiveness. The aim of this study is to evaluate the effect of a nutrition package on the treatment response and nutritional status of tuberculosis patients. Methods. We conducted a double-blinded placebo-controlled randomized trial in 360 pulmonary tuberculosis patients with concurrent diabetes or prediabetes. The participants were randomly assigned to receive a daily nutrition package (112 kcal, 9 g protein, and micronutrients) or a daily placebo package (112 kcal, 3 g protein, and no micronutrients) during tuberculosis treatment. The intervention lasted for six months. All participants received standard pulmonary tuberculosis treatment. The clinical symptoms, sputum smear, chest computed tomography, and nutritional status were monitored during the intervention. Results. The nutrition package improved the expectoration (intervention vs. placebo: 34.1% vs. 48.3% in week 1, 27.8% vs. 45.0% in week 2, 25.9% vs. 38.6% in week 3, 25.6% vs. 35.4% in month 1, 15.3% vs. 22.9% in month 2) and chest pain (2.3% vs. 9.0% in week 3, 3.6% vs. 8.3% in month 1, 4.3% vs. 10.0% in month 2, 1.8% vs. 6.4% in month 4). The nutrition package also increased hemoglobin, albumin, and lymphocyte counts. The nutrition package did not influence the sputum smear conversion in the whole population [hazard ratio (95% CI): 1.031 (0.685, 1.550), P = 0.885], but accelerated the conversion in patients without cavity [2.583 (1.180, 5.656), P = 0.018]. Conclusions. The nutrition package improved the clinical symptoms (e.g. chest pain and expectoration) and alleviated undernutrition (e.g. anemia and hypoproteinemia) among tuberculosis patients. The study was registered at the China Clinical Trial Registry Center (no. ChiCTR1900022294; https://www.chictr.org.cn).