The outputs of a principal component analysis (PCA) were ecologically interpreted, after the overall weed infestation indexes of all weed populations occurred in 82 sampling sites amounted to 57 5hm 2 cotton fields was assessed by visual scoring the level of weed infestation to cotton in seven scales. There was a close correlationship among the structure of weed community, crop rotation system and geographical factors. The marked difference in water content of soil resulted from the different crop rotation system with rice or dry crop significantly diversified weed distribution and occurrence in the cotton fields. The cropping rotation system with cotton rice resulted in identity of weed communities in the cotton fields. While soil and climate factors determined by geographical regions significantly influenced the communities in the cotton fields with succession of dry cropping system. All sampling sites were divided into four groups by PCA. The group of cotton cropping of rotation with rice had such dominants as Echinochloa crusgalli, Digitaria sanguinalis, Eclipta prostrata and Leptochloa chinensis. Three dry cropping succession groups respectively were Huaibei cotton growing region group with dominant weed Eleusine indica, Digitaria sanguinalis and Portulaca oleracea, Yanhai cotton growing group with Digitaria sanguinalis, Acalypha australis and Cephalanoplos segetum, and Yanjiang cotton growing group with Digitaria sanguinalis, Setaria viridis and Acalypha australis from the north to the south. According to the above weed distribution and occurrence pattern, weed control strategy was suggested.