In this paper, we studied the theory of personalized recommendation, compared the different recommendation technologies, and analyzed the applicability of several recommendation technologies to the O2O mode of e-commerce. And then an approach for mining personal context-aware preferences from the context-rich device logs, or context logs for short, and exploit these identified preferences for building personalized context-aware recommender systems was proposed. The experiment results show that the recommendation method put forward in the thesis is able to meet the demand of the O2O e-commerce mode.