To maintain a healthy diet, food and nutrition play a vital role. Everyone should keep up a healthy diet, but glycemic (diabetes) patients in particular should do so to maintain normal blood sugar levels. Nutrition therapy plays a significant role in preventing, managing, and controlling glycemic levels in blood for those patients. In this study, the recommendation of nourishment for diabetic patients has been carried out by utilizing clustering and optimization techniques. The patient’s profile has been clustered by utilizing improved K-Means algorithm. An algorithm called Improved Krill-Herd was used to increase the accuracy of recommendation outcomes. The performance of the system has been examined by various metrics such as Precision, Recall, F-measure, Accuracy, and Matthews Correlation. According to the findings, the suggested approach operates more effectively and generates advice to those diabetic patients that are optimized.