In this research, an efficient skin cancer detection technique based on Discrete Wavelet Transform (DWT) with Support Vector machine (SVM) is proposed. In this, all skin cancer images from ISIC 2018 dataset (International Skin Imaging Collaboration 2018 dataset), are pre-processed using a random sampling technique. The significant cancer features from these pre-processed images are extracted using DWT. This technique yields four significant features in the form of frequency sub-bands, namely High-Low (HL), Low-High (LH), Low-Low (LL) and High-High (HH). Here, the LL sub-band is having cardinal pixel information of an input image. So, the LL sub-band is further processed using SVM classifier to detect the type of skin cancer in dermatoscopic images. The detected results are compared with the existing results, for performing an evaluation using Accuracy, Precision, F1-score, Recall or Sensitivity and Specificity. The performance of our proposed method achieves 25% higher accuracy than the existing Random Forest (RF), K-Nearest Neighbor (KNN), Naïve Bayes and SVM classification.