Eye cancer is an uncommon disease. Despite its low incidence, ocular tumors such as uveal melanoma and retinoblastoma can be life–threatening in many cases. For small–to–medium ocular tumors, brachytherapy with radioactive 106Ru/106Rh plaques offers good outcomes in terms of local control and disease–free survival. Although these plaques have been used for decades, the measurement and calculation of the produced dose distributions remain challenging tasks. This thesis aims at improving the current knowledge on the theoretical and experimental dosimetry of 106Ru/106Rh plaques. First, we used Monte Carlo simulation of radiation transport to estimate accurate absorbed dose to water distributions produced by twelve plaque models. Secondly, we developed a practical experimental technique to measure in water the planar dose distributions produced by the plaques. For this experimental validation the EBT3 radiochromic film model was used which, in turn, required to characterize the absorbed–dose energy dependence of this new film model by means of Monte Carlo simulation. A detailed uncertainty analysis was carried out for simulation and experimental results. Both the simulated and measured absorbed dose distributions were compared with the available literature. The simulated dose distributions agreed within 3% with published data and with the EBT3 measurement results. Some discrepancies from the literature were also solved. This is the first, and currently unique, comprehensive dosimetry dataset for 106Ru/106Rh plaques. Therefore, it can become the basis for the first consensus dataset available. This dataset could also be useful for quality assurance of treatment planning systems. The simulation of the EBT3 film confirmed an improvement in the absorbed-dose energy dependence for this film model respect to older models. Applications for this film are found in many areas of medical physics. The measured dose distributions achieved a good agreement in absolute dosimetry with measurements by the manufacturer, and in relative dosimetry with the simulation results. The proposed method would allow users to check absolute dose distributions close to the plaque surface, and directly in water, thus avoiding the conversion from dose–to–plastic to dose–to–water, as it would occur if solid phantoms were used. With this method, a smaller experimental uncertainty was attained respect to the manufacturer and other authors.