In the Visco-Plastic Relaxation (VPR) strategy, the original elasto-plastic constitutive model is converted into a visco-plastic model with equivalent behavior for infinite time, and the classical Newton-Raphson iterations in the FEM analysis are replaced by time increments during which the stress points are returning progressively to the inviscid yield surface. Such strategy was developed for perfect visco-plasticity, and applied to the particular case of zero-thickness interfaces. The implementation was very much facilitated by the definition of a m-AGC tangent visco-plastic operator, which is sufficiently accurate to allow significant time steps without the need of iterations. In the present paper, the m-AGC tangent operator and the VPR strategies for the interface model are extended to the case of Hardening and Softening. This extension is based on a closed-form solution of the yield function evolution for given initial state and prescribed stress rate. The resulting m-AGC tangent operator constitutes an extension of the previous version for perfect visco-plasticity. This strategy is applied to the case of a well -established fracture-based interface model, and the methodology is verified with various academic examples representing different geo-mechanical scenarios. In these examples, it is verified that in the inviscid limit visco-plastic results match the elasto-plastic predictions, and that the VPR iterative strategy is capable of following the equilibrium path also during instabilities represented by a snap-back.