This paper presents a seamlessly integrated naturally fractured reservoir (NFR) modeling process for the ongoing Bati Raman heavy oil field rejuvenation project. As a part of the project, remedy of ongoing immiscible CO2 injection and successfully achievement of a near term steam flooding pilot test depend on understanding the heterogeneous natural fracture system. The naturally fractured rock and vug characteristics of the reefal/subreefal carbonate reservoir, and lateral and vertical facies distributions have been a challenge for establishing a successful 3D conformance process. An improved static model was constructed using the results of 2D and 3D seismic, sedimentological model and systematic fracture characterization. The new 3D seismic survey helped to identify the fracture regions, and had direct input to the new reservoir model with new structural information as well as fracture attributes. Fracture density and apertures of rare large true fractures were observed in the reservoir both on cores using a drill-core imaging system, and in FMI log findings. FMI logs provided the textures of vugs as well, ranging from less than 1 mm to 5 cm in size. The vuggy porous medium was handled in two ways: the interconnected vugs added into the fracture system since they build major pathways in the fractures between injector-producer pairs, while the isolated vugs are a part of the matrix system adjusted for compressibility. Matrix and fracture properties were distributed using sequential Gaussian simulation (SGS) and neural network algorithm constrained to the conventional log profiles within a 3D structural framework. This was followed by a calibration of well tests and CO2 breakthrough patterns, and hydrocarbon volume uncertainty assessment. The methodology optimizes a multi-disciplined integration and interpretation workflow for an unconventional multi-porosity complex model that is required in the distribution of flow properties such as saturation and pressure. Copyright 2007, Society of Petroleum Engineers.