Existence and uniqueness of solution of constitutive integration of non-associative plasticity has been an open issue until it was partly solved in Zheng et al. (2020), where the constitutive integration of elastic-perfect plasticity is reduced to a Mixed Complementarity Problem (MiCP), a special case of finite-dimensional variational inequalities, and the qualitative properties of the MiCP have been well established even for non-smooth yield surfaces and non-associated flow rules. The algorithm for the MiCP, called GSPC, has been proved globally convergent. In order to handle plasticity with hardening and softening behaviors, hardening functions are deemed the same position as stress components. In this way, hardening/softening plasticity reduces to elastic-perfect plasticity, and the GSPC is ideally suited to hardening/softening plasticity with no need to revise. The application of the proposed procedure to the Modified Cam-Clay plasticity is demonstrated. Comparisons are made with the return mapping (R-M) algorithm, indicating that those examples causing R-M to fail to converge can be easily solved using GSPC.