Objective:To build a hospital bed resource allocation model, for the reference of public hospitals in optimizing their bed resource allocation.Methods:Based on ReLU activation function, a hospital bed resource allocation model was constructed by combining DRG and public hospital performance appraisal requirements, including discharge person times, average length of stay, hospital bed utilization rate, proportion of surgery, proportion of fourth level surgery, case mix index, average bed day income and other indicators. When the existing number of hospital beds available was greater than the number of hospital beds allocated for the first time, a secondary allocation should be made. A tertiary general hospital was taken as an example for a model analysis.Results:As found in the model analysis, among the 2 729 beds of the hospital in the first allocation, 110 beds were left available for secondary allocation. The results of bed allocation of 40 inpatient departments in the hospital were as follows: 15 departments need more beds, 3 departments need more beds and shorter length of stay, 2 departments need no change, 1 department needs shorter length of stay, 4 departments need less beds, and 15 departments need less beds and shorter length of stay.Conclusions:The bed resource allocation model enriches the connotation of indicators, reflects the specialty characteristics. These indicators can be flexibly adjusted in combination with hospital development planning and budget management, hence conducive to refined management of hospital bed resources in public hospitals.