Abstract: In order to more scientifically solve the difficult problem of coal seam floor water inrush prediction and evaluation, first of all, a systematic and comprehensive review and comparison of the previous floor water inrush risk evaluation theories and methods, and in-depth and detailed analysis of the most advanced theory based on local variable weight According to the method of evaluating the water inrush vulnerability of the floor, it is considered that the method is not sufficient in the construction of the local state variable weight function, the variable weight interval and the determination of the weight adjustment parameters, and there is room for improvement. For this reason, a new three-interval local state variable weight function is constructed based on the existing theory and knowledge. The constant weight and the correlation coefficient of the constant weight are added to the state variable weight function, and the normalized value accumulation based on the main control factor is given. The new method for determining the threshold of the frequency variable weight interval is adopted, and the method of continuously adjusting and determining the final value of the weighting parameter based on the preliminary given empirical value is adopted according to the evaluation result. After that, a new three-interval variable weight model and its parameter determination method were used to evaluate the water inrush vulnerability of the Ordovician limestone floor in the study area, and the relatively vulnerable area was delineated. Finally, the weight adjustment law of the improved variable weight model is analyzed, and the limiting conditions of the constant weight correlation coefficient are determined through the sensitivity analysis method. It is found that a new three interval variable weight model with constant weight and constant weight correlation coefficient is added It not only reflects the core feature that the weight of the existing variable weight model changes with the change of the state value of the factor and its combination state, but also realizes the positive correlation between the degree of weight adjustment and the constant weight, which illustrates the impact of the variable weight vulnerability evaluation model this time. The improvement is scientific and feasible, and the research results are of great significance for improving the scientificity and practicability of the floor water inrush prediction and evaluation method.