Skin cancer, the most prevalent type of cancer around the globe, presents significant challenges in accurate diagnosis. Conventional diagnostic methods for skin cancer detection are time-consuming, costly, and often relies on the expertise of dermatologists. This research proposed a meta-learner-based model, utilizing deep learning models such as AlexNet, VGG-19, and ResNet18 to extract features and combine them with classifiers like SVM, Decision Tree, and Multi-SVM. The predictions obtained from different classifiers are input into a Meta Model, which averages predictions from base models and uses an ensemble technique to provide the final decision. The proposed is trained on a skin cancer dataset, showing superior performance with 95% accuracy, 96.9%-94.9% of precision, and 95% of F1-Score. This highlights the potential of meta-learner-based techniques in enhancing skin cancer detection.