Objective To investigate the relationship between nutritional risk status and implementation of nutrition therapy in mechanical ventilated (MV) chronic obstructive pulmonary disease (COPD) patients, so as to provide evidence for individualized nutrition therapy. Methods A prospective multicenter observational study was conducted. MV COPD patients admitted to Department of Intensive Care Units (ICU) of 10 County Hospitals in Zhejiang Province from January 2015 to January 2016 were enrolled, and according to nutrition risk screening 2002 (NRS2002) score, they were divided into nutritional high risk group (NRS2002 3-5) and nutritional extremely high risk group (NRS2002 6-7). Nutrition therapy situation and hospital mortality were compared between the two groups; multivariate Cox regression analysis was used to analyze the risk factors affecting the prognosis of patients with COPD under mechanical ventilation. Kaplan-Meier curve was used to analyze the prognosis at 30 days; receiver operating characteristic (ROC) curve was used to test the robustness of multivariable regression analysis. Results ① One hundred and six COPD patients with MV were analyzed; among them, 90 patients were in the nutritional high risk group, and 16 were in the nutritional extremely high risk group. There were no significant differences in age, gender and body mass index (BMI) between the two groups (all P > 0.05); the acute physiology and chronic health evaluation Ⅱ (APACHEⅡ) score, NRS2002 score in patients of nutrition risk extremely high group were obviously higher than that in patients with nutrition high risk group (APACHEⅡ: 24.9±6.1 vs. 20.3±5.8, NRS2002 score: 6.3±0.5 vs. 4.2±0.8, both P < 0.05). ② Patients in both groups received early enteral nutrition (EN) therapy, the proportion of patients in nutritional extremely high risk group received early EN was lower than that of patients in the nutritional high risk group [12.5% (2/16) vs. 17.7% (16/90)], along with the prolongation of hospital stay, the proportions of patients beginning to receive the EN were gradually increased in the nutrition extremely high risk group and high risk group, after 2 days the EN increased significantly, and reached the highest value on day 6 after entering ICU [100.0% (16/16), 98.9% (89/90), respectively]; within 3 days after admission into ICU, the proportion of EN in nutrition extremely high risk group was obviously lower than that in nutrition high risk group, and from day 4, there was no statistical significant difference in proportion of EN between the two groups (all P > 0.05). The time to start parenteral nutrition (PN) treatment was relatively early admission to the ICU on day 1 and the proportion of this therapy was high in the two groups [56.2% (9/16), 27.7% (25/90), respectively], the PN proportion did not decrease with the length of hospitalization and the increase of EN. The proportion of patients in the nutrition extremely high risk group who started PN treatment was higher, which reached 56.2% admission to the ICU on day 1.③ With extension of hospital stay, the calories of EN were gradually increased in the nutritional high risk group, the highest calories in nutritional high risk groups was 4 318 (3 912, 4 812) kJ/d at day 7; while the highest calories in nutritional extremely high risk groups was 3 602 (2 167, 4 615) kJ/d at day 6 and a slight decreased at day 7; the difference of calories within the first week between the two groups had no significance (all P > 0.05). The calorific value of PN therapy remained at a constant level during hospitalization within 7 days, and after admission into ICU for 4-5 days, the target range of calories was achieved. ④ Kaplan-Meier survival curve analysis showed that the mortality at 30 days in the extremely high risk group was significantly higher than that in the high risk group [62.5% (10/16) vs. 11.1% (10/90), χ2 = 15.4, P < 0.01]. ⑤ Multiple cox-regression analysis showed that NRS2002 scoring was the independent risk factor affecting the mortality of patients in hospital [odds ratio (OR) = 2.08, 95% confidence interval (95%CI) = 1.39-3.12, P = 0.005]. ⑥ ROC curve analysis: according to ROC curve analysis of the effectiveness of multi-factor regression model, area under ROC curve (AUC) was 0.79, sensitivity was 70.00%, specificity was 74.42%, positive likelihood ratio was 2.74, negative likelihood ratio was 0.40, 95% confidence interval (95%CI) was 0.702-0.864, P = 0.001, and it showed that the regression model had a good prediction effect. Conclusions MV COPD patients have significant nutritional risk and all receive early EN therapy. The proportion of beginning to use PN treatment in patients with nutritional extremely high risk is relatively high. Initial nutritional status is the independent risk factor of poor prognosis in MV patients with COPD.