Forecasting of underground pressure for working surface (UPWS) is very important in mining technology industry for safety production. These factors (roof lithologic, mining height of coal seam, geological structure, etc.) directly affect the difficulty of forecasting trends in this field. The original Grey Model is not hard to understand and very simple to calculate, with satisfactory accuracy, but it is also lack of flexibility to adjust the model to acquire higher forecasting precision. This paper studies feasibility and effectiveness of a novel Grey model together with the concept of the Bernoulli differential equation in ordinary differential equation. In this research, we name this newly proposed model as Nonlinear Grey Bernoulli Model (NGBM). The NGBM is nonlinear differential equation with power index r. By controlling r, the curvature of the solution curve could be adjusted to fit the result of one time accumulated generating operation of raw data. NGBM is then applied in UPWS forecasting. The results of long terms forecasting prove the novel NGBM is feasible and efficient.