Information visualization facilitates a viewer to quickly digest information in massive data. Many computer vision and pattern recognition problems may be posed as the analysis of a set of similarities between objects. For many types of data, the theoretical structure is supposed to be circular or spherical, and therefore cannot be mapped in a Euclidean space. We aim to propose an efficient numerical method for embedding data on circle or sphere based on similarity data. We first give a model for spherical embedding. Then, making use of the concept of Euclidean distance matrix, we propose a method to obtain embedded points. At last, we apply our method to a variety of data and draw several maps which can show relations among objects. In each case the embedding maintains the local structure of the data while placing the points in a metric space.