To comprehensively analyze the wedge stability whose parameters have various uncertainties, a method was proposed to analyze the stability of wedge probability based on blind data theory and Fisher optimal segmentation method. A blind data evaluation model of wedge stability considering a variety of wedge parameter uncertainties was created to determine the wedge stability. The evaluation model of the blind number of the stability of the wedge was given, and the model was applied to the classical double-wedge stabilization, in which the result was compared with that from the traditional mean safety coefficient, numerical simulation, and Monte Carlo sampling (MCS). The compared results show that the result of evaluation model of wedge stability based on blind data theory and Fisher optimal segmentation method is coincident with the generally accepted MCS method, which verifies the reliability and efficiency of the proposed method. The proposed method was applied to the stability evaluation of the wedge in Xingcheng to provide data support for the prevention and control of the wedge.