Abstract: Prediction of water outburst in seam floor of mines is a complex theoretic and technical problem, which has characters of fuzziness, grey and incompatibility.By building the factor index system and getting nine main factors through the grey correlation, the contribution weights are ascertained by the entropy adjustment.According to the grey matter-element analysis theory, it builds the linear grey matter-element and optimal grey matter-element model respectively, the discrimination system of combined grey matter-element for water outburst in seam floor of deep mines is exploited by VB 6.0.The two module in parallel operation and mutual verification can much raise the result precision, it is basically capable of forecasting deep seam mining faces which are all threatened by the Ordovician limestone water.The answer of simulation analysis applying this system to the realistic water-bursting cases is in good accordance with the truth condition.The system will be of some values for controlling and preventing deep seam water-bursting from coal floor in complex state.