MRI-alone based radiation treatment planning (RTP) is an emerging field of research where the main concern is to perform definition of structures of interest and dose calculations primarily based on MR images. However, MR images lack the necessary electron density information, thus, it has to be assigned to it, and this is done by estimating a CT image from the available MRI data. In this paper, we propose a method to predict a pseudo-CT image from standard T1-weighted MRI data using a multi-metric image registration combined with a multi-atlas fusion technique. We used an atlas database consisting of T1-weighted MRI and CT brain images and employed multi-metric deformable image registration to capture the target image anatomy and the final pseudo-CT is created by fusing the initial deformed CT atlas set. The generated pseudo-CT is compared to the original CT using different evaluation metric. Furthermore, we compared our method to the state of the art single atlas approach. Results show that the proposed approach can predict a reliable CT image from MRI data that can replace the original CT for RTP