Snowmelt runoff is themain composition ofwater resources inmiddle and high latitude region. Research on the snowmelt runoff process are mi portant in monitoring and foresting flood and drought disaster as well as managing water resources. In recent years the studies on distributed snowmelt runoff model are based on the application and local mi provementof foreign mature models such as the SRM, have more lmi itations in their applicabilities. A Distributed SnowmeltRunoffModelbased on RS and GIS has been designed and explored in this paper. Themodel consists of three parts, grid snowmeltmode,l grid snowmelt flow generationmodel and grid snowmelt flow concentrationmode.l Itwasbased on energy balance andwaterquantitybalance. The grid snowmeltmodelwasbased on the energybalance. And the grid flow modelwas based on thewater volume balance aswell as the distributed grid runoffmode,l thus has formed a completely,distributed snowmelt runoffmodelwhich could apply to the actual service. Snowmelt and runoff processwere all based on grid scale, allofthe input/outputdatawere distributed processed, the prmi ary formatwereTIFF, so themodelwasprovided with distributing character and had strict physicalmechanism. An mi portant concept of“Unit Period ofTmi e”was put forward in the paper, and thereby developed a“Degree and CentSnowmeltRunoffModel”from the traditional“Degree-Day Method”which haswidespread application nowadays, for itwas not fit for the practical applications because of its glancing tmi e-resolution, in the snowmelt period, the processes were usually very unabiding, so the tmi e-step in day is not appropriate, butwith the“Degree and Cent SnowmeltRunoffModel”, the computing tmi e-stride of the snowmelt flood precaution decision support system could adjust itsel,f and tomake use in the real-tmi e computation and the forecast. Italso brought forward another mi portant concept of“Freezing andMelting Coefficient”, which was ami ed to explain the snow freezing andmelting repetitive physicalmechanism and this concept is significant to exactly hold the snowmelt process,because thewatermelting of snow/ice usually refreeze in the afternoon. Besides, aDistributed SnowmeltRunoffStmi ulating System byGIS has been developed, which is the platform and support of the running of the Distributed SnowmeltRunoff Model, has allofthe basic functionsofGIS, such as spatialanalysis, hydrologicalanalysis, dynamic smi ulation, and so on.Both themodel and the system are the kernelmodules of SnowmeltFlood Prewarning Decision Support System. At last,based on a greatdealofremote sensingdata, obtained distributed inputdata for themode,l forexample, distilled snow cover informationwithUnmixing PixelMethod, and land surface temperature (LST) with Split-window Arithmetic designed by ZhihaoQin, KebiaoMao and so on, with ourproper correction. Basic geo information data likeDEM and its relevantdata such as shope and asect, field smi ultaneous observation data mainly including weather data and hydrology data, were disposedwith uniform spatial resolution (30m×30m in thispaper) and the same data format(TIFF), stmi ulated the runoff process in representative snowmeltperiod in 2006 ofthe typicalstudy area, Juntanghu riverbasin, a closewatershed, which located in themiddle ofthe northernTianshanMountain,Xinjiang, China, the area of it is833·57km2, with theDistributed Snowmelt Runoff Stmi ulating System, the tmi e-step was 15 minutes. An average relative error of this model with the synchronous truemeasuring data was under 0·18, indicating that themodelwas able tomeet the need of snowmelt flood alarming and foresting.