Bacillus thuringiensis (Bt), belonging to the Bacillus cereus (Bc) group, is commonly used as a biopesticide worldwide due to its ability to produce insecticidal crystals during sporulation. The use of Bt, especially subspecies aizawai and kurstaki, to control pests such as Lepidoptera, generally involves spraying mixtures containing spores and crystals on crops intended for human consumption. Recent studies have suggested that the consumption of commercial Bt strains may be responsible for foodborne outbreaks (FBOs). However, its genetic proximity to Bc strains has hindered the development of routine tests to discriminate Bt from other Bc, especially Bacillus cereus sensu stricto (Bc ss), well known for its involvement in FBOs. Here, to develop tools for the detection and the discrimination of Bt in food, we carried out a genome-wide association study (GWAS) on 286 complete genomes of Bc group strains to identify and validate in silico new molecular markers specific to different Bt subtypes. The analyses led to the determination and the in silico validation of 128 molecular markers specific to Bt, its subspecies aizawai, kurstaki and four previously described proximity clusters associated with these subspecies. We developed a command line tool based on a 14-marker workflow, to carry out a computational search for Bt-related markers from a putative Bc genome, thereby facilitating the detection of Bt of interest for food safety, especially in the context of FBOs.