Granite fractured basement reservoir contributes higher 40% of the world's oil and gas reserves. However, geological modeling of fractured reservoirs is complex and presents unique challenges in comparison with conventional reservoirs. It is extremely difficult to achieve the best results for a future development plan. This research presented the new workflow to enhance the accuracy of porosity and permeability models for a fractured reservoir in offshore Vietnam by using Artificial Neural Network (ANN) and co-kriging method. ANN was employed to solve problems that conventional modeling has not been successful. The seismic attributes selection was used for initial ANN generation. Then, the prediction property model was established through ANN training process. Well log data was used for correlation to cross-validation the predictive models. Next, the co-kriging algorithm was created the porosity and permeability models. Also, the Drill Stem Test (DST) data was used for history matching models to confirm the Co-kriging approach. The history matching was iterated until the geological model achieved the best matching with DST data. The history match shown the excellent fitting between simulation model and measurement data. Overall, we conclude that ANN and co-kriging are useful method for developing reliable workflow in fracture basement reservoir.