Abstract: Meteorological data presents a subst antial challenge for scientific visualization application, because of the very nature of the data itself. The visualization of meteorological data is the last element in the pipeline of meteorological data processing. It enables the forecaster or the researcher to understand the complex circumstances by representing the spatial distributions and temporal evolution of weather situation, and allows scientists to extract knowledge efficiently from rapidly expanding multidimensional datasets. With the amount of data increasing exponentially, a fundamental question is "What is the most efficient and most easily underst andable way to examine these datasets?" It is difficult to use traditional methods designed for two dimensional data sets to extractscientific information from these multidimensional data sets. In this paper, a summarization of graphical representations of multidimensional data sets in meteorological area was presented, a powerful meteorological datavisualization system running on PC-Windows operating system was introduced, several typical weather (cyclone, typhoon, etc.) analysis of mult idimensional data sets generated by numerical model were used to demonstrate some of the capability of this system. Although the visualization systems have played an important role in the understanding of complex weather situations, but there are still many limitations in current visualization system. For example, too many complex parameter adjusting are discom modious for the forecaster, can't support the experience or knowledge of the forecaster which formed in past analysis of weather situations (viz.intelligent analysis), there is a lack of ability to sumup the conceptive models of special weather situation which are very important in practice, and can't have access to the fullrange of meteorological data, etc. At the end of this paper, these limitations, and expectations in improving were indicated.