Video segmentation is a first step towards video analysis. Video segmentation is responsible for identifying scene boundaries, logically connected frames. Scene is a fundamental part of the video because it captures actions in a form of coherent sequence of frames. Although, many algorithms for scene cut detection were proposed, in most approaches only one visual feature was used. Here we propose a novel algorithm for scene cut detection that utilizes multiple visual features including color histogram, edge histogram and frame wavelet statistics. Experiments run on complex video data prove algorithm effectiveness compared to other scene cut detection algorithms.