This paper presents an innovative NPSHr prediction procedure. Derived through analyzing the definition of NPSHr and the characteristics of pump head vs. NPSH curve (i.e. head drop curve), this procedure can significantly reduce the number of simulations needed in NPSHr prediction. For the majority of cases, NPSHr can be predicted with reasonable accuracy for one flow rate in as little as three simulation runs. This procedure is also very robust since it can avoid running simulations under severe cavitation conditions. Therefore simulation convergence is improved and simulation time is reduced for each simulation step. In the paper, the methodology of the proposed approach is derived and explained. The complete procedure with suggested strategies is laid out in detail. Then the procedure is demonstrated against an industrial pump case. NPSHr prediction results are compared with experimental data. Important factors which can affect NPSHr prediction are also indentified and discussed.