Faba bean (Vicia faba) is a widely adapted nitrogen fixing grain legume that produces seeds with high protein content. It is an agriculturally attractive crop with a good disease resistance profile, but targeted breeding is hampered by the lack of a reference genome and gene set. Here we develop a faba bean gene expression atlas and pan-transcriptome. The pan-transcriptome was generated using data from four different accessions (Hedin, Hiverna, 153b, 2378), including data from both aerial and root tissues. The quality and utility of the pan-transcriptome was assessed by mapping reads from accessions that had not contributed data to the pan-transcriptome. Here, the high mapping rates suggested that the pan-transcriptome represents a comprehensive faba bean reference gene set and that it will be a useful resource for differential gene expression analyses and genome annotation. In addition, the gene expression atlas provides valuable information to support gene function discovery.