Considering the sparsity of traditional collaborative filtering recommendation algorithms in electronic commerce systems, a new collaborative filtering algorithms based on enhanced similarity and implicit trust is presented. Firstly, a new method based on JMSD and user's preference to compute the similarity measure is presented. Secondly, a method to compute the direct trust fused with the interactive experience is proposed. Then, a method to compute the implicit trust based on direct trust and trust propagation is presented. Finally, this paper presents a model to compute the rating predictions based on the enhanced similarity and implicit trust. The experimental results in Movielens and Epinions show that the new algorithm improves recommendation quality in MAE and coverage.