Thyroid-related disorders represent a growing global health concern, affecting individuals across all age groups and often remaining undetected during early stages. Timely identification of thyroid abnormalities plays a critical role in preventing long-term complications and improving patient outcomes. This paper presents an intelligent, web-based thyroid health detection system that integrates machine learning techniques with automated data extraction and user-centric healthcare services. The proposed system analyzes thyroid hormone values—Thyroid Stimulating Hormone (TSH), (T3), and (T4)—along with demographic attributes and self-reported symptoms to identify thyroid conditions such as hypothyroidism, hyperthyroidism, goiter, and normal thyroid function. A Random Forest classifier is employed as the core predictive model due to its robustness and high accuracy with structured medical data. Optical Character Recognition (OCR) is incorporated to extract hormone values directly from uploaded laboratory reports, reducing manual effort and potential data entry errors. The system offers dedicated dashboards for patients and doctors, enabling report analysis, appointment scheduling, health tracking, and secure communication. Experimental evaluation demonstrates that the proposed approach achieves reliable diagnostic performance while maintaining usability and accessibility. The system serves as a practical, non-invasive, and cost-effective decision support tool for preliminary thyroid assessment [1].