Target detection and recognition is the ultimate aim in computer intelligence. Given an example of linear target, bridge over water in SAR image, the algorithm of denoise and segmentation is designed based on fuzzy theory, the detection of bridge edge is realized by minimum and Radon transform, and a measurement of judging parallel is proposed in this paper. In the process of designing algorithm, some problems have been considered and solved, including: (a) restrain furthest the multiplicative noise and enforce the feature of target; (b) obtain the area of bridge over water in complicated background; (c) get the edge of bridge, namely extraction of linear feature; (d) the final work in detection and recognition of bridge such as parallel judgment. The experimental results show that the algorithm is effective