The risk prediction in dengue patients and risk criteria definition for dengue disease are included in the range of challenges that have been addressed. For life-saving measures, surveillance and clinical care, it is essential to effectively and accurately predict the risk stage for dengue disease. The existing studies suggested several techniques but it met a few of the shortcomings in terms of computational and time complexities with prediction accuracy. Hence, this study proposed a novel approach by applying machine-learning algorithms for dengue disease prediction. Initially, the patient’s data like Age, Sex, Body Temp, Platelets Level Count, Electrolytes Count, Red Blood Cell (RBC) Count and White Blood Cell (WBC) count are gathered from the Government Medical College, Bellary. After that, this study proposed a novel Coati Optimization algorithm-based Support Vector Machine (CO-SVM) model to predict the presence or absence of dengue diseases based on the symptoms of Fever, Headache, Rash skin, Severe Muscle and joint pain, Severe pain behind the eyes, Swollen Glands and Exhaustion. For this study, the patient’s data were collected from Government Medical College, Bellary during the time period of 2019 to 2022. The total of 1480 patients is admitted in hospital. Out of this, 978 patients are dengue positive cases and the rest of 502 cases are dengue negative classes. Statistical parameters like prediction accuracy (A), Mean Squared Error (MSE), Mean Absolute Error (MAE) Root Mean Square Error (RMSE) measures are to validate the effectiveness of the proposed CO-SVM method for dengue disease prediction and the proposed work outcomes are compared to other existing state-of-art methods.