A great deal of computational power and time is necessary for simulating highly heterogeneous fractured reservoirs with complex geometry; the efficiency of these computations is a major subject especially in large-scale heterogeneous and fractured reservoirs. So, simulating large scale, complex and fractured reservoirs with both minimum time and maximum accuracy is the scope of current work. The BiCG-Stabilized solver preconditioned by CPR-AAMG has been developed to achieve acceptable results of high efficiency and robustness for large heterogeneous fractured Black-Oil models. The solver's efficiency is demonstrated in an Iranian fractured field model with heterogeneity. As an observation, the NF preconditioner which is embedded in ECLIPSE has low efficiency in simulating giant fractured reservoirs. On the other hand, the advantageous efficiency of the proposed CPR-AAMG preconditioning technique increases with the complexity of the model. The findings indicated that the CPR based preconditioners have better performance than classical ILU based preconditioners according to the number of linear solver iterations in RETINA. CPR-ILU0 is the most efficient preconditioner, based on speed and calculation cost (numbers of linear solver iterations). RETINA can take larger time steps in comparison with ECLIPSE, and therefore has better elapsed time. Comparison with the results of other preconditioners developed in ECLIPSE approved that the developed CPR-ILU0 in RETINA is robust. The main conclusion would be that RETINA has an exceptional and stable performance for simulating highly heterogeneous fractured models. The novelty of the proposed CPR-AAMG preconditioner is the ability to solve three dimensional black-oil models of giant, complex and fractured reservoirs with reduction in time and required computational power.