Multi-modal registration is a fundamental step for many medical imaging procedures. In this paper, the sum of conditional variance (SCV) similarity measure is proposed for 3D multi-modal medical image registration. The SCV similarity measure is based on minimizing the sum of conditional variances that are calculated using the joint histogram of the two images to be registered. Standard Gauss-Newton optimization is used to automatically minimize this measure which allows fast computational time and high accuracy. Experimental results show that our proposed approach is robust, computationally efficient and also more accurate when compared with the standard mutual information (MI) based approach and also the recently proposed sum-of-squared-difference on entropy images (eSSD) approach.