Objective To investigate the current associated factors of dietary knowledge, intake and nutritional status in malignancy inpatients, and the malnutrition causes involved in dietary nutrition knowledge level, behavior, providing recommendations to patients for nutrition education and intervention. Methods 535 participants, from 18 hospitals were investigated by a self-made questionnaire composed of dietary knowledge and behavior. Physicians asked and recorded the level of dietary intake and appetite scoring of the participants. Nutritional risk screened with NRS 2002 and dietary survey by 24h dietary recalls were completed by a dietitian. Besides, the target energy intake and the target protein intake was calculated by “rule of thumb” recommended by ESPEN guideline, comparing the difference between the actual intake and target intake. Results According to the questionnaire, 95.2% of participants thought it was important to have a good dietetic habit, and nearly half of them searched for guides on how to diet; 70% of the patients had no clear idea of what was a scientific dietary; 82% of patients had contradictary dietary knowledge contradiction; 64.2% of patients would listen to the opinion of the attending physician when contradiction happened. The three main ways to learn about diet are attending physician, network, and TV, respectively 26.0%, 18.5% and 16.1%. Importantly, 99.6% of the patient mistakes on dietary knowledge, for example, crab, chicken, lamb, fish and prawns should not be eaten by cancer patients. In addition, more than 90% of participants taking ganoderma lucidum spore powder, sea cucumber, ginseng, cordyceps sinensis and so on. 93% of the patients never received a qualified nutrition education. Besides, 15.6% of the participants had nutritional risk (NRS 2002≥3 scores). The actual daily energy intake was 1169.20±465.97kcal, which was significantly less than target energy intake (P<0.01), amounting to 65.3% of the target requirements. Actual daily protein intake was 46.55±21.40g, which was significantly less than target energy intake (P<0.01), amounting to 74.4% of target requirements. On the other hand, 69% of the participants were “Not too bad, Ok, Good or Very good” according to the records of physicians, while 34% of them did not reach 60% of the target requirements through dietary recall. So the results of physicians’ inquiries did not reflect the actual dietary intake for cancer patients. Conclusions The survey indicated that cancer patients had poor understanding of the scientific dietary nutrition and were in low level of normative nutritional education among Chinese malignancy inpatients. Dietary intake of most cancer patients decreased, and the actual intake cannot be responded by NRS 2002 scoring or the physicians’ inquiring. It is necessary to enhance the cooperation between dietitians and physicians, and hold targeted nutrition education, in order to improve the level of dietary knowledge.