This project aims to develop a machine learning-based system for the early diagnosis of vector-borne diseases using patient symptom data. The dataset used comprises 64 binary features representing the presence or absence of symptoms such as fever, headache, vomiting, and muscle pain. The target variable, “Prognosis,” is categorical and includes diseases like Lyme disease, Zika, and Rift Valley fever. The dataset was preprocessed using several techniques: missing values were imputed using column means; categorical labels were encoded numerically; and features were standardized using StandardScaler. In order to reduce dimensionality while preserving 95% of the variance, Principal Component Analysis (PCA) was employed. To tackle class imbalance, SMOTE was used to oversample the minority classes. The data was split into training and testing sets with an 80:20 ratio. Various models—including Random Forest, Logistic Regression, Decision Trees, Support Vector Machines, and Multi-Layer Perceptron Classifiers—are being applied. Model effectiveness will be evaluated using metrics such as accuracy, precision, recall, F1-score, and ROC-AUC, all derived from the confusion matrix. While no definitive results are available yet, this approach is designed to assess and compare model effectiveness in accurately predicting diseases, with the ultimate goal of aiding early medical intervention and reducing global health burdens.