[Purpose/significance] This paper aims to identify personalized reading needs of university readers,predict its development trend,and realize personalized push services. [Method/process] By analyzing the feasibility of the small data of university readers applying to library personalized intelligent services,libraries push personalized intelligent services combining with small data acquisition,preprocessing,detection and prediction of personalized reading needs with decision recommendation mechanism. [Result/conclusion] Results show that the small data of university readers can accurately reflect their personalized reading and knowledge needs,and provide support for making the decision and constructing the model about personalized intelligent service by library. Under the big data environment,this research can provide some theoretical guidance and practical value.