Based on the experimental results of uniaxial ratchetting-fatigue interaction obtained at room temperature for the materials presenting different cyclic softening/hardening features [Kang GZ, Liu YJ, Li Z. Experimental study on ratchetting-fatigue interaction of SS304 stainless steel in uniaxial cyclic stressing. Mater Sci Eng A 2006;435-436:396; Kang GZ, Liu YJ. Uniaxial ratchetting and low-cycle fatigue failure of the steel with cyclic stabilizing or softening feature. Mater Sci Eng A, in press], two kinds of stress-based fatigue failure models were proposed to predict the fatigue life of the materials by addressing the ratchetting-fatigue interaction occurred. In the primary stress-based failure model (namely, SBF model), the effect of ratchetting strain produced in the asymmetrical cyclic stressing on fatigue life was reflected by an item dependent upon the stress ratio R. Comparison with the experimental results shows that the SBF model provides a good prediction for the cyclic stable material, annealed 42CrMo steel, but over-estimation for the materials presenting cyclic softening feature, i.e., SS304 stainless steel and tempered 42CrMo steel, since the effect of cyclic softening/hardening feature on the ratchetting-fatigue interaction is not concluded in the primary model. To overcome such shortcoming, the primary model was then extended by introducing a new variable FP to address the effect of cyclic softening feature. In the modified stress-based failure model (namely, MSBF model), the variable FP was determined by applied maximum stress and stress ratio as well as the cyclic softening feature of the material. It is shown that the MSBF model gives fairly good predictions for both SS304 stainless steel and tempered 42CrMo steel by addressing the effect of cyclic softening feature on the fatigue life reasonably.