Cystic fibrosis is a human monogenic genetic disease caused by mutations in the cystic fibrosis (CF) gene, which encodes a membrane protein which functions as a channel: the cystic fibrosis transmembrane conductance regulator (CFTR) protein. The most frequent mutation, a deletion of phenylalanine F508 (delta F508), is located in the first nucleotide binding domain of CFTR: NBF1. This mutation leads to a folding defect in NBF1, responsible for an incomplete maturation of CFTR. The absence of CFTR at the surface of epithelial cells causes the disease. Determination of the three-dimensional (3D) structure of NBF1 is a key step to understanding the alterations induced by the mutation. In the absence of any experimental data, we have chosen to build a 3D model for NBF1. This model was built by homology modelling starting from F1-ATPase, the only protein of known 3D structure in the ATP binding cassette (ABC) family. This new model defines the central and critical position of F508, predicted in the hydrophobic core of NBF1. F508 indeed could be involved in hydrophobic interactions to ensure a correct folding pathway. Moreover, this model enables the localization of the LSGGQ sequence (a highly conserved sequence in the ABC family) in a loop, at the surface of the protein. This reinforces the hypothesis of its role for mediation of domain-domain interactions of functional significance for the channel regulation. Finally, the model also allows redefinition of the ends of NBF1 within the CFTR sequence. These extremities are defined by the secondary structure elements that are involved in the NBF1 fold. They lead to reconsideration of the C-terminal limit which was initially defined by the end of exon 12.