This exploratory study serves two purposes: a) identify the institutions (organizations, programs, regulations, data infrastructures, etc.) named in some major International Environmental Law conventions and in the decisions or resolutions of the associated Conferences of the Parties (prior to 2015) and b) evaluate the potential contribution of Natural Language Processing-here SPACY and CoreNLP-in their Named Entity Recognition (NER) function, to this identification in an extensive body of legal texts. After describing the main steps of computer-based NER and the performance criteria of the algorithms (recall and precision), we present the results of the analysis of about a thousand legal texts constituting our corpus. We combine the use of NER tools and the manual screening of raw results, a list of more than 800 institutions involved in environmental governance is established, 110 of them being more specifically involved in the governance of oceans and marine resources. The retrieval and classification of declarative information extracted from the corresponding websites then makes it possible to provide a picture of the governance of the oceans seen through these conventions, but also to specify the contributions and limits of the computer-based NER. The application of this method necessitates an interdisciplinary collaboration.