Abstract: Netlike texture is very common in our daily life. However, it is very difficult to be described and detected because of its distinct characteristics. This paper chooses nets in sports game as research object and presents a netlike texture detection method of using fractal texture features and wavelet transform. First, multi-scale wavelet decomposition is employed on video frames; then fractal features of multi-resolution images are obtained. A group of fractal features are chosen as feature vector according to causality theory and an SVM classifier is used to recognize the net in images. The experiments show that this method is robust and effective under the condition of intense distortion and complicated background. This method can be extended to other netlike texture detections.