The aim of this study is to show how a numerical database constructed using <q>uncertainty quantification techniques</q> can be a useful tool for the analysis of the tsunamigenic potential of a seismic zone. This methodology mainly relies on the construction and validation of some <q>emulators</q>, or <q>meta-models</q>, used instead of the original models in order to evaluate and quantify the uncertainty and the sensitivity of the tsunamis height at a given location to the seismic source parameters. The proposed approach was tested by building a numerical database of nearly 50 000 tsunamis scenarios generated by the Azores-Gibraltar Fracture Zone (AGFZ) and potentially impacting the French Atlantic Coast. This seismic area was chosen as test-case because of its complexity and the large uncertainty related to its characterization of tsunamigenic earthquake sources. Finally, the tsunami hazard resulting from uncertainty quantification was presented and discussed with respect to the results which can be obtained with a more classical deterministic (or scenario-based) approach. It must be underlined that the results from this study are the illustration of a general methodology through a case study with simplified hypothesis, which is not an operational assessment of tsunami hazard along the French Atlantic Coast.