Abstract: In order to explore the relationship between cotton seedling emergence and soil environment under "dry sowing and wet germination" conditions,the soil hydrothermal gas and seedling emergence data of 34cotton fields under "dry sowing and wet germination" in Shaya County,South Xinjiang were investigated,and analyzed the correlation between seedling emergence rate and soil temperature,soil water content,electrical conductivity and soil compactness by using partial correlation method to clarify the degree of influence of each factor on seedling emergence rate,and established a comprehensive prediction model for cotton seedling emergence rate based on soil temperature,soil water content and soil compactness.A comprehensive model for predicting cotton seedling emergence based on soil temperature,soil water content and soil consolidation was established.The results showed that the correlation between soil habitat and seedling emergence rate was as follows:soil water content＜ soil bulkiness ＜ soil electrical conductivity ＜ soil temperature ＜ maximum diurnal temperature difference.Soil slump,soil electrical conductiving,soil temperature and seedling emergence showed a linear relationship,and water content and seedling emergence showed a quadratic function relationship,the soil moisture content during the seedling stage was 19.03%,with the highest seedling rate.Two multivariate nonlinear models were developed for seedling emergence with soil bulkiness,soil cumulative temperature and water content,and one of the models was selected to predict cotton seedling emergence,with soil cumulative temperature being the most sensitive to seedling emergence and soil water content and soil bulkiness being the second most sensitive.The accuracy of both the single factor model and the comprehensive factor prediction model were complex to the simulation accuracy requirements,but the multivariate nonlinear model of the comprehensive factor was superior to the other three single factor models.In conclusion,the established multivariate nonlinear model can be used to guide the actual production of "dry sowing and wet germination" cotton fields in Southern Xinjiang.