This study proposes a lightweight model based on support vector machines (SVM) for the diagnosis of Parkinson's disease, which is convenient for migration embedding. It is primarily applied in community and elderly care settings. The method achieves accurate Parkinson's disease diagnosis through the recognition of hand-drawn spiral lines. The aim is to increase awareness of geriatric syndromes like Parkinson's and expand the coverage of diagnostic services, providing more timely medical intervention and care for the elderly population. The study collects hand-drawn spiral line datasets from both Parkinson's patients and healthy participants. SVM is employed as the diagnostic model, leveraging its adaptability to high-dimensional data and nonlinear relationships. During the training process, the model considers handwriting features, including pressure and angles. In the testing phase, the system rapidly and accurately identifies Parkinson's disease symptoms by recognizing handwritten patterns. The research results indicate that the SVM-based Parkinson's rapid diagnostic method achieves a correct rate of 93.3%, demonstrating promising applications in community and elderly care facilities.