Background: Capturing the healthfulness of mixed dishes per serving can be a challenge. Objectives: To develop a Nutrient Density Calculator (NDC) for mixed dishes to guide recipe innovation and reformulation. Methods: Recipe files and nutrient composition data for mixed dishes came from the US Department of Agriculture Food and Nutrient Database for Dietary Studies (FNDDS 2021–2023). Included were mixed dishes with meat, poultry, seafood, grains, vegetables or legumes as their main ingredient, Asian and Mexican dishes, pizza, sandwiches, burgers, and nuts and seeds (n = 1570). Observed servings as consumed came from the National Health and Nutrition Examination Survey data (NHANES 2017–2020/2021 –2023). The NDC version 643 had three components: NR6 based on six nutrients to encourage (protein, fiber, calcium, iron, potassium, vitamin D); MP4 based on ingredients (fruits and vegetables, whole grains, pulses, nuts and seeds); and LIM based on three nutrients to limit (saturated fat, added or total sugar, sodium). An abbreviated version, NDC343, used NR3 (protein, fiber, calcium). Alternative NDC models tested different component weights. Nutrient density was calculated per serving. Results: Both NDC643 and NDC343 gave higher scores to vegetables, pulses, and nuts and lower scores to meat-based dishes. Equal weighting of the three components produced NDC scores that were correlated with Nutri-Score (r = −0.64 for NDC343 and −0.70 for NDC643) and Health Star Rating values (r = −0.74 for NDC343 and −0.75 for NDC643). Correlations with energy density were −0.25 for NDC343 and −0.33 for NDC643. The NDC643 model was sensitive to recipe improvements for a variety of mixed dishes—more so than the Nutri-Score or the Health Star Rating. Conclusion: The three-component NDC model, combining both nutrients and ingredients, successfully captured nutrient density of mixed dishes and was sensitive to recipe improvements.