Reservoir simulation is a powerful tool to model fluid flow within oil and gas reservoirs and predict their behaviour. This dissertation is devoted primarily to model some thermal enhanced oil recovery (TEOR) methods. Two software were used for this purpose and namely; Comsol Multiphysics® and CMG® (Computer Modelling Group). The dissertation can be classified into three parts and all of them are standalone that discuss different topics within TEOR. The work starts with reviewing enhanced oil recovery (EOR) methods with concentration on thermal methods (TEOR) for heavy oil and bitumen. Basics of mathematical modelling of single, two-phase, and three-phase flow in porous media that is the base of all commercial and scientific reservoir simulation software are reviewed. Formulations of the set of representative PDEs are reviewed and other formulations are suggested and applied directly in subsequent sections in Comsol Multiphysics®. Part-1: The application of finite element method (FEM) in reservoir simulation has been discussed and evaluated using Comsol Multiphysics package which is based on Galerkin approach. In the demonstrated problems, the mathematical model is solved using mathematics module in Comsol Multiphysics. Energy equation in 1D, Buckley-Leverett benchmark, two-phase flow model on ¼ inverted 5-spot scheme in 3D, and SAGD process PDE model are all solved and discussed. FEM using Comsol Multiphysics looks promising at moderate mobility ratios. Part-2: A comparison of water flooding with steam injection in heavy oil reservoirs as secondary stage is demonstrated and discussed. The whole modelling was achieved by CMG-STARS. A comparison of five different scenarios is shown. SPE4 comparative project data were used for this purpose. The results showed that steam can achieve more recovery in a short period of time with an ultimate recovery factor higher than cold recovery followed by steam flooding process. Part-3: A series of flooding and in-situ combustion experimental work that has been achieved in Kazan Federal University in cooperation with Institute of Drilling Engineering and Fluid Mining (IBF) is elaborated briefly and discussed. Four experiments with different core samples (consolidated and unconsolidated) were run between 05-2020 and 05-2021. The samples were taken from a Russian extra-heavy oilfield with initial viscosity around 600,000 cP. The results were evaluated and a numerical model was built using CMG-STARS. The numerical results were correlating the experimental results. Relative permeability data were history matched for flooding processes and this data was used for in-situ combustion model. Modelling of the reactions in in-situ combustion was a challenge to match the experimental results. The final results showed that steam injection was not the best recovery method for this oilfield and in-situ combustion was the best available technique with the highest recovery factor.