This paper proposes a computationally-efficient electromagnetic (EM)-computational fluid dynamics (CFD) coupling approach for a water-cooled Interior Permanent Magnet (IPM) motor. The numerical simulation of multiple fluids and their interaction with solid parts can be challenging. The proposed approach relies on the heat transfer coefficient (HTC) decomposition of different fluids/coolants inside the machine to generate an HTC look-up table (LUT) as a function of the coolant inlet flow rate. The HTC-LUT is then utilized as a surrogate model for the stationary coolant to decouple the fluid-to-fluid interaction and, hence, expedite the iterative approach. This reduces the computational time by almost two-thirds as compared to the multiple fluid approach while preserving the fidelity of the model. Additionally, the proposed approach formulates the correlation between the rotor speed, coolant flow rate, convective heat transfer coefficients, and the temperature rise, particularly for hot-spot locations in the end-windings. This approach uses a two-dimensional EM model coupled with a three-dimensional fractional CFD model. Thus, it reduces the computational cost and retains the model simplicity. The viability of the proposed coupling approach is also validated through lumped parameter thermal network (LPTN) analytical approach. The results from both approaches under different cooling conditions, flow rates, and current densities are in a good agreement.