Objective To determine the reliability and validity of the Duke Social Support Index (DSSI)-23 scale in a rural elderly population. Methods A stratified cluster random sampling strategy was adopted to select 702 rural participants aged 60 years or over in Guangyuan. Those with severe hearing disorders or mental problems were excluded. The reliability and validity of DSSI-23 scale were evaluated using internal consistency, split-half reliability，content validity, convergent validity, discriminant validity and structure validity. Results The DSSI-23 had a Cronbach’s α efficient of 0.881 and 0.918 split-half reliability. Item-scale correlation coefficients exceed 0.35 （P 2/df=6.884， CFI=0.807，GFI=0.850，RMSEA=0.092. Conclusion DSSI-23 scale has good reliability in the elderly population, but unsatisfactory validity．Further adjustments of the scale are needed.
