Time dimension has always been an important measurement scale of human society, producing large amounts of time-series data all the time in the fields of science, engineering, economics and so on. Exploring correlation features and periodic trends of multi-dimensional time-series data is the emphasis in the research of visual analytics. A visual interactive system named TimeSprial is proposed on the basis of past cases and visualization methods in this paper. The system is designed based on the concepts of time granularity and time primitive, so as to explore correlation features and periodic trends of data dimensions through visual analysis. TimeSprial integrates various visualization layout methods of periodic data, such as ring diagram and curve graph, assisted by a variety of interactive models. Finally, case analysis of actual data sets shows the effectiveness of our approach in the exploration and understanding of multi-dimensional time-series data.