This research looks into the application of Support Vector Machines (SVMs) for image classification within the medical field, focusing on kidney cyst detection in CT-scanned images. It investigates the effectiveness of SVMs with different kernels in distinguishing between images containing cysts and those depicting normal kidneys. This research seeks to develop and evaluate the performance of SVM models using linear and polynomial kernels to classify kidney cysts in CT scan images. However, the presence of non-linear patterns in the data can pose challenges for traditional machine learning algorithms. The models carried out are trained on a dataset of CT-scanned images labeled as cysts or normal. A comparison using performance metrics, including accuracy, precision, recall, and F1 score is done to determine the most effective model for cyst classification. The linear kernel SVM achieved an accuracy of 86.90%, demonstrating good performance in classifying both cyst and normal images, whereas the polynomial kernel SVM outperformed the linear model, achieving a higher accuracy of 89.74%. Our findings suggest that the polynomial kernel SVM outperforms the linear kernel SVM for classifying kidney cysts in CT-scanned images. The polynomial kernel effectively captures non-linear patterns, leading to improved accuracy and maintaining efficiency. The model can help a professional in the diagnosis of an individual by keeping its own classification in front.