Early detection of malnutrition in hospitalized patients is of paramount importance. As poor food intake is a marker of malnutrition risk, a simple and accurate method to monitor intake is valuable. This quality assurance project aimed to determine if food service workers (FSW) were able to provide accurate estimates of patient intakes through visually assessing meal trays at an acute care hospital. FSW conducted visual estimates of patient trays after meals using the meal plate pictorial rating scale adapted from the My Meal Intake Tool and translated their estimates into one of 5 consumption levels (0%, 25%, 50%, 75%, or 100%). A total of 401 patient meal estimates were validated using the food weighing method. Spearman’s correlations between percent calories consumed (determined by weight) and estimates by FSW were 0.624 (n = 137, P < 0.001), 0.771 (n = 134, P < 0.001), and 0.829 (n = 130, P < 0.001), for breakfast, lunch, and supper, respectively. Paired Wilcoxon tests and the Kruskal–Wallis H test showed that accuracy varied for breakfast, lunch, and supper. The overall sensitivity and specificity of FSW for detecting patient intake ≤50% was 81% and 88%, respectively. These findings identify that FSW can accurately estimate patient intake, contributing an important marker for the detection of malnutrition.