Recommendation systems take advantage of products and users information in order to propose items to targeted consumers. Collaborative recommendation systems, content-based recommendation systems and a few hybrid systems have been developed. We propose a dynamic and adaptive framework to overcome the usual issues of nowadays systems. We present a method based on adaptation in time in order to provide recommendations in phase with the present instant. The system includes a dynamic adaptation to enhance the accuracy of rating predictions by applying a new similarity measure. We did several experiments on films data from Vodkaster, showing that systems incorporating dynamic adaptation improve significantly the quality of recommendations compared to static ones.