Abstract: Traditional recommendation algorithm not only ignores the connection between the users’ distance, but also ignores the different evaluation standard. For this problem, this paper proposes a based on distance attenuation and score trends improved collaborative recommendation algorithm，which use the distance decay function and the score trends algorithm to improve the collaborative algorithm. Experimental results show that the recommendation algorithm can not only improve the accuracy, but also enhancing the efficiency of commodity recommendation system

Key words: distance, rating trends, collaborative recommendation, commodity recommendation