Résumé : Forecasting the global energy demand is remarkably important for future energy policy and security. Considering various scenarios for world energy production and demand, the role of natural gas in shaping future energy demand will be notable, derived by its environmental advantages and versatility relative to other combustible fuels. This work is in the context of Combustion for Low Emission Application of Natural Gas project (CLEAN-Gas) funded by European Union's Horizon 2020 research and innovation programme under Marie Sklodowska-Curie Innovative Training Network (ITN), aiming to propose an innovative approach to improve natural gas combustion in industrial processes including detailed chemistry and computational fluid dynamics. Towards the goal, this work aims to extend the knowledge on hidden aspects of natural gas in the low environmental impacts combustion technologies through the development of comprehensive, detailed kinetic mechanism with predictive capabilities in a wide range of operating conditions of interest for real systems. The kinetic mechanism of natural gas (C1-C3 fuels) describing the oxidation and combustion of natural gas is conceived and developed in a modular and hierarchical approach. This thesis is an effort to fill the pressing need of a reliable and widely validated kinetic mechanism, specially developed for modern combustion systems with near-zero emission. More than 200 different experiments containing more than 6000 data points of various apparatuses such as plug flow reactor (PFR), Jet stirred reactor (JSR), shock tube, and 1-D laminar flame are collected from literature for the sake of extensive and critical model comparisons. This database is the most extensive set of experimental data available, which is beneficial for understanding the complex combustion processes of modern combustion technologies that have been hindered from successful integration into the industry. A systematic study is performed on the combustion characteristics, less-known aspects and critical reaction pathways involved in MILD and oxy-fuel combustions. Diluent effects are evaluated in detail, and it is noteworthy to highlight that physical and chemical effects of diluent on the reactivity, laminar flame speed, ignition delay time, and formation of products are strongly dependent on the operating conditions (temperature, pressure, and equivalence ratio). Therefore, the analysis of the dilution effects is very case sensitive, and the contribution of each characteristic may vary accordingly. Both H2O and CO2 dilution reduces the system reactivity. The effect of H2O is more notable due to chemical effects related to enhanced collisional efficiencies at the operating conditions of JSR experiments. On the contrary, CO2 has a higher impact on inhibiting flame propagation at higher temperatures, mainly due to the thermal and radical scavenging effects. Moreover, the effect of CO2 addition on methane ignition delay times is very marginal. The most likely explanation is that CO2 is scarcely reactive during the ignition, because of its stability, and does not actively modify the pool of radicals. Finally, despite the satisfactory model prediction and reasonable agreement shown, the underlying impact of rate parameters uncertainty on model prediction is still not negligible. It has been shown that even if the kinetic mechanism is complete and free of any missing reaction pathway, rate coefficient uncertainties generally precludes the possibility of predicting relevant combustion properties in some peculiar conditions. One of the primary reasons has been the lack of general agreement as to how to move forward in obtaining a comprehensive, unified and predictive model for fuels combustion.