The safety and risk assessment of CO 2 storage in geological formations requires a robust and iterative methodology based on an objective assessment, which shall provide an analysis and assessment of potential risks to health, safety and environment. The application of this methodology from the initial stages of the project will facilitate achieving its objectives. The results of the methodology should be twofold: the quality of the site from the point of view of the risks and the associated uncertainties. In the early stages of a project involving scarcely known natural systems, the methodology should take into account the unavoidable uncertainties in the available information and its impact on the risks, through a formalized quantification of those. In these phases the models used are mainly qualitative. As the project progresses and more information is available, the risk assessment methodology should allow gradual and continuous transition from qualitative data based models to quantitative ones.Taking all these into account, in this work are presented the methodologies commonly used, based on those developed and fine-tuning for the past 20 or 30 years to the study of Deep Geological Repositories of high-level nuclear wastes, as well as the development carried out to estimate the risks of Hontomín Technological Development Plant, implemented under the formalism of Bayesian Networks (BNs).