<![CDATA[AZY Clinic experiences obstacles in the process of predicting diseases, especially diabetes, which still takes time and has not been supported by an early prediction system. This study aims to implement a hybrid algorithm of Random Forest and Support Vector Machine (SVM) for web-based diabetes prediction using the Software Development Life Cycle (SDLC) Iterative Model. The research methods include data collection, preprocessing, dataset sharing, model training, evaluation, and system implementation. The dataset consists of 768 data from the Kaggle public dataset for model formation and testing and 60 AZY Clinic patient data for the period April to June 2026 as system implementation data. The hybrid model uses Random Forest as a base classifier to produce prediction probability values which are then combined with four normalization attributes, namely age, blood pressure, sugar levels, and BMI, as a stacked feature for the Support Vector Machine as a meta classifier. The test results showed that the hybrid model obtained an accuracy of 61.04%, precision of 45.83%, recall of 39.29%, F1-score of 42.31%, and AUC of 0.5904. The model was implemented in a web-based application using Laravel, Python, and Flask to generate predictions of diabetes status and patient risk levels. The implementation of the system can support the initial prediction process and assist medical personnel in obtaining risk information based on patient data as supporting information, not as a substitute for medical personnel assessment]]>
