Support Vector Machines (SVM) are a machine learning technique that has been used for segmentation and classification of medical images, including segmentation of white matter hyper-intensities (WMH). Current approaches using SVM for WMH segmentation extract features from the brain and classify these followed by complex post-processing steps to remove false positives. The method presented in this paper combines the use of domain knowledge, advanced pre-processing (based on tissue segmentation and atlas propagation) and SVM classification to obtain efficient and accurate WMH segmentation. Features generated from up to four MR modalities (T1-w, T2-w, PD and FLAIR), differing neighbourhood sizes and the use of multi-scale features were compared. We found that although using all 4 modalities gave the best overall classification (average Dice scores of 0.54 ± 0.12, 0.72 ± 0.06 and 0.82 ± 0.06 respectively for small, moderate and severe lesion loads, using 3x3x3 neighbourhood intensity features); this was not significantly different (p = 0.50) from using just T1-w and FLAIR sequences (Dice scores of 0.52 ± 0.13, 0.71 ± 0.08 and 0.81 ± 0.07 for the same lesion loads and feature type). Furthermore, there was a negligible difference between using 5x5x5 and 3x3x3 features (p = 0.93). Finally, we show that careful consideration of features and preprocessing techniques leads to more efficient classification which outperforms the one based on all features with post-processing, and also saves storage space and computation time.