Water quality is a crucial factor in environmental health, human well-being, and the sustainability of natural ecosystems. Traditional methods for water quality assessment, while accurate, often required laborious and time-consuming laboratory analyses that controlled large-scale and real-time monitoring. This research paper presents a novel approach to predict water quality assessment using machine learning techniques applied to images captured by smartphones. Our model uses a custom-built dataset of water bodies to classify water based on visual characteristics into four categories: clean water, muddy water, polluted water, and water contaminated by algae. We also developed a web-based platform that allows users to upload water images, which are then analyzed in real time by our train model to provide instant classification. This system democratizes water quality monitoring, making it accessible to a wide audience and enabling more frequent and decentralized data collection. Our study demonstrates the feasibility of using smartphone images for water quality assessment, offering a scalable and efficient solution for real-time environmental monitoring, with potential applications in sustainable water resource management.