Banana is an important export-oriented fruit in South China.Division of banana for climate suitability provides scientific basis for rational use of natural resources and optimization of banana distribution.To overcome the shortcomings of traditional methods of climate division that over-rely on knowledge and experience and have no ability to learn,a approach based on a so-called ID3 decision tree method is provided.Annual mean temperature,annual precipitation,accumulated temperature of ≥10 ℃,annual extreme minimum temperature,number of consecutive days with average temperature of ≤ 8 ℃,and mean annual sunshine duration are taken as climatic factors to build Grids Reckoning Models.Climate data from 1971 to 2000 and geographical information data of 90 weather stations in Guangxi are used to build models,with each of the climate factors projected as l km×l km raster data and discretization operation carried out on each of the climate factors based on an index of climatic suitability division.Combined with discreted raster data with fieldwork data,samples of banana for climate suitability division are used to build a decision tree based on the ID3 algorithm and the total classification accuracy is 92.5%.Results shows that the ID3 decision tree method is suitable for banana climate division.