Deep learning has led to impressive performance on a variety of object detection tasks recently. However， it is rarely applied in ship detection of synthetic aperture radar (SAR) images. This paper aims to introduce a detector based on deep learning into this field. We analyze the advantages of the state oftheart Faster R-CNN detector in computer vision and limitations in our specific domain. Given this analysis, we propose a dataset and four strategies to improve the detection result. The dataset contains ships in various environments, such as image resolution, ship size, sea condition, and sensor type. It can be a benchmark for researchers to evaluate their algorithms. The strategies include feature concatenation, transfer learning, loss function optimization method, and other implementation details. We conduct some comparison and ablation experiments on our dataset. The result shows that our proposed method obtains better accuracy and higher efficiency.