This paper aims to improve user searching experience by optimizing the query suggestions on social reading network. Firstly, a user behavior feature database (library) is constructed by using standardized co-occurrence matrix based on query words; this behavior library is integrated into the socialized reading platform query suggestions list, from which other historical searching information is introduced to optimize the effect of the query suggestions. Then, through the simulation of query prompt process, the richness and the recall rate of prompt effect are quantified separately. The results...