Objective To evaluate the new mini nutritional assessment-short form (MNA-SF) in the assessment of nutritional status in the elderly inpatients with chronic diseases. Methods A total of 2 861 elderly patients aged 65 or older were selected in this study, who were hospitalized in Chengdu Elderly Care Hospital from March 2016 to May 2018. All the patients were assessed with the MNA-SF within 48 h after admission. Cronbach coefficient α was used to evaluate reliability of MNA-SF, and Kaiser-Meyer-Olkin (KMO) and Bartlett spherical test to evaluate its validity. According to MNA-SF score, the patients were divided into normal nutrition group (n=319), malnutrition risk group (n=1 046), and malnutrition group (n=1 496). The three groups were compared in nutrition-related indicators, and the nutritional status was predicted with the receiver operating characteristic (ROC) curve. SPSS statistics 22.0 was used for analysis. ANOVA, LSD-t test, rank sum test or χ² test was used for comparison.Results The incidence of malnutrition was 52.3% (1 496/2 861) in the inpatients with chronic diseases. Cronbach′s alpha of the MNA-SF was 0.711, showing a good reliability. KMO value was 0.827, indicating that factor analysis could be performed. Bartlett′s spherical test χ² was 27.616 (P<0.05), and the spherical hypothesis was rejected, indicating that factor analysis was possible. The eigenvalues of seven MNA items[dietary change, weight loss in recent three months, activity ability, stress or acute disease, neuropsychiatric disease, body mass index (BMI), calf circumference (CC)] were all greater than 1, and the cumulative variance contribution rate was 83.14%, indicating a reasonable validity of MNA-SF. The age of normal nutrition group, malnutrition risk group and malnutrition group tended to increase, while BMI, CC, albumin, prealbumin, transferrin and hemoglobin tended to decrease, with statistical significance (P<0.05). ROC curve showed that the best cut-off point for predicting nutritional status of the elderly inpatients with chronic diseases was 10 points, with a sensitivity of 78.93%, a specificity of 83.59%, and an area under curve of 0.797 (95%CI 0.781-0.811, P<0.05). Conclusion MNA-SF is applicable for nutritional assessment of the elderly inpatients with chronic diseases. With a high incidence of malnutrition, nutrition intervention in the elderly patients with chronic diseases should be implemented as soon as possible.