The “Training and development dataset for information extraction in plant epidemiomonitoring” is the annotation set of the “Corpus for the epidemiomonitoring of plant”. The annotations include seven entity types (e.g. species, locations, disease), their normalisation by the NCBI taxonomy and GeoNames and binary (seven) and ternary relationships. The annotations refer to character positions within the documents of the corpus. The annotation guidelines give their definitions and representative examples. Both datasets are intended for the training and validation of information extraction methods.