Blood vessel segmentation ia an important step in image analysis for tumor angiogenesis study etc. With the purpose of extracting line structures of blood vessels, some filter-based methods are used to segment vessels. However, the design of accurate and automatic vessel segmentation algorithms is still challenging, due to the variety and complexity of images, especially in brain blood vessel segmentation. In this work, we address a problem of automatic segmentation of brain micro-vessels structures in light-sheet image stacks. To segment micro-vessels in large-scale image data, we propose a method using convolutional neural networks (CNN) as a pixel binary classifer Three convoutonal layers and one fully connected layer are used in th CNN model. We extract a patch of size 32x32 pixels in each acquired brain vessel image stack as trainning data set to feed into CNN for classification. This network is trained to output the probabilty that the center pixel of input patch belong to vessel structures. To build the CNN arhitecture, a series of mouse brain vascular images acquired from a commercial light sheet fluorescence microscopy (LSFM) system were used for training the model. The preliminary experimental results demonstrated that our approach is a promising method for effectively extracted micro-vessels structures in brain images.