It would be faster to copy and paste each reference from the pdf into an EndNote record.
It might be possible to import the references if the data is restructured in a format where each field (e.g., author name, year, title, etc.) is tagged so the information can be placed in the corresponding EndNote fields. It can be done but takes time to do so (e.g, copying the information into an Excel spreadsheet, where each column represents a single EndNote field).
Or using a proprietary database. I know Scopus and probably Web of Science have a way to select and export cited references of a record in their databases.