Osteoarthritis, a prevalent degenerative joint disease, significantly impairs quality of life, particularly among the elderly. Traditional diagnostic methods often involve invasive and expensive imaging techniques. This project aims to develop a non-invasive, real-time prediction system for osteoarthritis using the K-Nearest Neighbors (KNN) algorithm, a robust machine learning approach. The core of this system is its ability to accurately and comprehensively collect sensor data from the user's joints. The system integrates a variety of non-invasive sensors—flex sensors, MPU6050 sensors, and piezoelectric sensors—interfaced with a NodeMCU microcontroller. These sensors offer high sensitivity, fast response time, and durability, making them ideal for capturing critical data relevant to osteoarthritis detection. The data from these sensors is transmitted to the cloud and analysed using the KNN algorithm to predict the likelihood of osteoarthritis. The dataset, sourced from Kaggle, is split into 70% for training and 30% for testing. The KNN algorithm is applied to classify individuals into different osteoarthritis risk categories. This non-invasive, portable, and efficient solution offers a promising alternative to traditional diagnostic methods, making osteoarthritis prediction more accessible and cost-effective.