Abstract: Some pattern recognition methods including dynamic clustering analysis, minimal spanning tree and principal component analysis have been used to recognize the extents of both lattice-match and thermo-matchat300KforthesamplescontainingZnOsemiconductormaterial. Fromthepatternrecognitionanal-ysis, it is concluded that there exists a lower lattice and thermal misfitsbetween ZnOand threeⅢ-Ⅴgroup nitrides. The results alsoshow that pattern recognition analysisseems to be asuitable tool toselect substrate or bufferlayerforepitaxygrowth, aswellastodesignoptoelectronicdevicesofblueandvioletlightsemicon-ductor materials, such as ZnO and GaN.