Dual Energy X-ray Absorptiometry (DXA)-based finite element (FE) modelling has emerged as a potential tool for better assessment of osteoporotic hip fracture risk. Automation of this complex and computationally-intense procedure is the prime requirement for its clinical applicability. The aim of this study was to develop a fully automatic DXA-based finite element tool and assess its discrimination ability and short-term repeatability. The proximal femur was automatically segmented from clinical hip DXA scan and the subject-specific FE model was constructed for simulating sideways fall. Hip fracture risk indices (HFRIs) were calculated using two ways (along a femur cross-section and over a region of interest, ROI). Hip fracture discriminability increased when moved from femur cross-section based to ROI based HFRI calculation. A significant increase in hip fracture discriminability from baseline femoral neck and total hip bone mineral density (BMD) was achieved with ROI based HFRIs. Promising short-term repeatability was observed for HFRIs (coefficient of variation, CV, 3~3.5%). After removing representative poor cases, CVs were less than 3%. These preliminary results establish the potential of the proposed automatic tool for hip fracture risk assessment and justify large-scale clinical evaluation of its ability to predict incident hip fractures.