Aimed at solving the difficulties, such as low efficiency and limited exploration range encountered in finding groundwater with the traditional methods, a new method was presented by using remote sensing technology in this paper. Based on multi-spectral data (ETM data) and spatial data (SRTM data), a forecasting model was built to produce a probability rating map for finding shallow groundwater in the arid and semi-arid areas. According to investigations, a conclusion is drawn that the results of the model are satisfied, which have been testified by the later geophysical exploration and drilling. Thus, the model can serve as a guide for finding groundwater in the arid and semi-arid regions.remote sensing, shallow groundwater, forecasting model, multi-spectral data, spatial data How important the water is! However, many countries are suffering from lack of surface water, particularly in the arid and semi-arid areas. Therefore, people have to exploit groundwater in those regions. But, unlike surface water, groundwater is very difficult to explore. For a long time, people used traditional methods to explore groundwater, like drilling or geophysical exploration, but these methods have many limitations. Take drilling for an example, the method is quite expensive and has low efficiency. In this situation, an advanced technology for exploring groundwater is needed. With the development of the modern science and technology, remote sensing is such an advanced method. At present, there are only four methods for exploring groundwater by using remote sensing technology: hydrologic geology analysis [1] , environmental analysis [2] , anomaly information of surface temperature [3] and quantitative inversion model [4] . However, these methods ignore the influence of nature, such as terrain or geology. Therefore, sometimes, the methods cannot obtain satisfactory results. In order to solve the problem, this paper presents a new method for finding groundwater.
