Abstract: 【Purpose/significance】 In order to improve the subjectivity and fuzziness of traditional evaluation and better reflect the de‐
mands of users,an evaluation system of digital resource service quality of academic libraries is constructed on the basis of the indexes
that users contribute,and using grey correlation analysis.GA-BPNN model is proposed to evaluate the service quality of digital resourc‐
es of academic libraries【. Method/process】The paper consults users' opinions from the perspective of their demands,extracts the key elements that affect the evaluation of digital resource service quality in academic libraries,and builds a relatively scientific and reason‐able hypothetical model by means of empirical analysis.Then,samples are collected and the results of service quality evaluation are represented by grey correlation degree.Finally,MATLAB R2019b is used to conduct simulation experiments,the performance of GA?BPNN model and BPNN model are compared and analyzed.【Results/conclusions】The actual output of GA-BPNN is close to user's
subjective evaluation results,and the accuracy and stability of GA-BPNN are significantly better than the BPNN model,which reflects
the feasibility and rationality of the evaluation model.It is expected to provide some references for the research on digital resources con‐struction,service quality evaluation and service innovation of academic libraries and other libraries【. Innovation/limitation】The paper puts forward a method to evaluate academic libraries’digital resource service quality based on user demand.The main limitation is that the coverage of survey data is not comprehensive enough.