Quantitative computed tomography (QCT) -based finite element analysis is a commonly accepted approach for prediction of mechanical behavior of bones. The objective of this research is to suggest linear criterion in order to accelerate and increase the precision of predicting of failure load in femoral bone. Accordingly, ten fresh frozen femora were QCT scanned and performed to use in this study. The specimens were loaded under eight different orientations. Finite element model for these samples were generated from QCT images, and related mechanical properties were calculated for each single voxel based on the value of density. In addition, the models were analyzed by linear finite element method. Risk factor that defines as the strain energy density divided to yield strain energy for each element was used for calculations of failure load. These values were sorted for particular loads in finite element model, and the correlations between experimental and numerical results were compared. Finally, eight linear criterions for eight different load conditions were presented which shows magnificent correlation between empirical results (average slope: 0.8903 and average R2: 0.8668). These correlations make it possible to accelerate the prediction of femoral fracture load in various orientations. This research shows a robust and fast method for prediction of failure in bones that can be used for multiple loads and orientations.