Various human pathologies, such as cancers or, notably, exudative age-related macular degeneration (AMD), are characterized by an abnormal vascular proliferation (angiogenesis), which in the case of this type of AMD, can rapidly lead to blindness if no countermeasures are taken. In the past few years, it has become possible to stop - or at least to notably slow down - the evolution of AMD, by applying treatments based on photodynamic therapy (PDT), whereby, through the action of a molecule acting as photosensitizer (PS) which is targeted preferentially to the choroidal neovessels (CNV), the effects of this abnormal angiogenesis are stopped by provoking the closure of the neovessels. In particular, the PDT group at EPF-Lausanne has, in the past, already tested on the model of the chick embryo chorioallantoic membrane (known as the CAM model) several PSs of potential interest as substitute to the product currently considered as standard (Visudyne® , itself co-developed in this group) for the treatment of the CNV associated with AMD. The aim of this thesis was to study new molecules and/or formulations, potentially suitable as substitutes to the current photodynamic treatment of AMD, while diminishing its negative side-effects: general phototoxicity due to diffusion of the product out of the vasculature and further lack of selectivity leading to (undesirable) partial occlusion of the normal vasculature. In order to enable extrapolation and comparison of the results obtained, we have focused our attention on vascular effects only and administered the tested PSs under essentially similar conditions (formulation in DMSO; same delivery system through biodegradable nanoparticles built with the same polymer). In addition, the same animal model (the CAM model) was used throughout the entire study, the formulations being administered by intravenous injection. This model, importantly, allows for an easy and real-time semi-quantitative visualisation and recording of the vascular effects of a PS induced during a PDT session. In the first part of this work, we demonstrate that the tetra-phenylporphyrin molecule, when formulated in nanoparticles based on poly(D, L - lactic acid, of well defined size (ca. 200 nm or less) is a very interesting PS-candidate potentially suitable as a replacement of the currently used one. We show that inclusion of this PS in nanoparticles notably reduces the PDT's general phototoxicity. We believe therefore that this PS would deserve more thorough pre-clinical studies, on an animal model closer to the human case (for example, Ryan's model). The second part of this thesis aimed at elaborating a strategy allowing, in due time, for an improved selectivity of occlusion of the neovasculature, said strategy aiming at an active targeting of neovessels. Recently, a number of possible biochemical targets have been identified, including the so-called ανβ3 receptors, which are highly expressed at the level of the neovessels' endothelial cells. Our approach consisted in coupling a fluorescent molecule (fluorescein, easily visualized by epi-microscopy) to a protein (a kistrin) which contains the RGD peptide sequence, known for its affinity for the ανβ3 receptors. We have thus been able to demonstrate that the kistrin was effectively linked to these receptors and, therefore, that active targeting of the neovessels had indeed taken place. A complete proof of this concept by realizing the coupling of this kistrin with a PS molecule will be the next step in this research.