For the computational simulation of ductile fracture, most computational methods suffer from mesh dependent issues, and thus a regularization term and/or a length scale parameter is introduced to alleviate the mesh dependency. Additionally, the crack surface representation is essential in relation with the mass and energy conservation. For the investigation of the ductile fracture simulation, a computational framework named as Gurson-Cohesive Model (GCM) for simulating 3D ductile fracture is presented associated with void growth and coalescence phenomena [1]. The proposed model idealizes the process of ductile fracture as continuum damage evolution, cohesive crack initiation, nonlinear softening along crack surface, and complete failure. The Gurson model is employed to describe continuum damage, while the cohesive zone model is utilized to introduce discontinuous cracks. The transition between the Gurson and cohesive zone models is taken into account systematically using a porosity-based crack initiation criterion in conjunction with the stress triaxiality effect. The proposed model, i.e., the unified computational framework, is validated for two ductile alloys, 15-5 PH steel and 316 stainless steel. The computational results successfully reproduced the experimental results of uniaxial tension tests and compact tension tests of various thicknesses and pre-crack lengths. Additionally, the proposed model predicts local behavior such as deformation measured by DIC and 3D scanning techniques as well as the length of crack propagation and the depth of crack tunneling according to the crack opening displacement. Furthermore, the results show strong and stable convergence under mesh refinement without the aid of any regularization term and/or length scale.