Background: Prevalence of malnutrition in hospitals has been reported around 20% and increases during hospitalization. The “Rate-a-Plate” method has been developed to monitor dietary intake and identify patients whose nutrition status deteriorates during hospitalization, but has not yet been validated. The objective was to study the validity and reliability of the method (phase 1) and redesign and revalidate a revised version (phase 2). Methods: Detailed food records provided a reference method. A priori difference of >20% in energy or protein between the reference and the “Rate-a-Plate” method was determined as clinically relevant. Intraclass correlation coefficients were used to determine the reliability. Results: In phase 1, 24 patients were included with a total 67 test days. In phase 2, 14 patients were included, 28 test days. In phase 1, the “Rate-a-Plate” method underestimated intake by 422 kcal (29%, ICC 0.349, 95% CI 304–541) and 5.7 g protein (10%, ICC 0.511, 95% CI 0.0–11.5). Underestimation was found in 65% and 23% for energy and protein intake, respectively. Underestimation was higher when patients had higher intake. In phase 2, underestimation was 109 kcal (7%, ICC 0.788, 95% CI −273 to 56) and 3.7 g protein (6%, ICC 0.905, 95% CI −8.4 to 1.0). In 32% and 21% of the cases, energy and protein intake were underestimated. Conclusion: The revised version of the “Rate-a-Plate” method is a valid method to monitor energy and protein intake of hospitalized patients and can be filled out by nutrition assistants. A larger validation study is required.