We have recently proposed an original 2D multi-resolution algorithm for solving a TLM-like formulation of the Indoor propagation problem. This approach is referred to as MR-FDPF (multi-resolution frequency domain parflow). In this paper, a fast and simple extension for predicting radio coverage in multifloored building is proposed. Because a full 3D is time consuming, three 2.5D schemes are investigated and compared. The accuracy of MR-FDPF is firstly assessed in a single floor configuration. Results exhibit a RMSE of about 3.5 dB, after an automatic calibration. Then 2.5D extensions are evaluated on a multifloored building. The more sophisticated one yields a RMSE of about 4.4 dB for the inter-level prediction. The potential of this approach and the obvious limitations when compared to a true 3D approach are discussed.