Registration is an important task in image processing used to match two or more images. In the field of medical imaging, registration is used to match anatomic structures from two or more images taken at different times to track for example the evolution of a disease. Intensity based techniques are widely used in multimodal registration. To have the best matching, a cost function expressing the similarity between these images is maximized. To resolve this problem, we evaluate two global optimization techniques that are genetic algorithms and simulated annealing. In this paper we show some results obtained in medical images registration.