<![CDATA[Diabetes mellitus is a chronic metabolic condition that requires proper dietary management to keep blood glucose levels stable. Fruit is recommended for people with diabetes as it contains important nutrients; however, differences in carbohydrate and fibre content mean that careful consideration must be given to the choice of fruit combinations and portion sizes. This study aims to apply the Mamdani fuzzy logic inference method to generate recommendations for fruit combination portions based on carbohydrate and fibre content. Secondary data were obtained from the Indonesian Food Composition Table (TKPI), which comprises 10 types of fruit. Data processing involved fuzzification, the formulation of a rule base, Mamdani inference using the Min-Max operator, and defuzzification using the centroid method. The defuzzification results yield a compatibility value which is used to determine the recommendation category and the appropriate fruit combination portion. This method has been implemented in a web-based application to provide automatic recommendations to users. The research findings indicate that the Mamdani fuzzy logic inference method can effectively process carbohydrate and fibre content into measurable and easily understandable recommendations regarding fruit combination portions. The proposed system can assist people with diabetes in selecting suitable fruit combinations and appropriate portion sizes, thereby supporting healthier dietary planning and facilitating daily decision-making regarding fruit consumption.]]>
