Cystic fibrosis (CF) is a genetic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) gene. The 2789+5G>A CFTR mutation is a quite frequent defect causing an aberrant splicing and a non-functional CFTR protein. Here we used a CRISPR adenine base editing (ABE) approach to correct the mutation in the absence of DNA double-strand breaks (DSB). To select the strategy, we developed a minigene cellular model reproducing the 2789+5G>A splicing defect. We obtained up to 70% editing in the minigene model by adapting the ABE to the PAM sequence optimal for targeting 2789+5G>A with a SpCas9-NG (NG-ABE). Nonetheless, the on-target base correction was accompanied by secondary (bystander) A-to-G conversions in nearby nucleotides, which affected the wild-type CFTR splicing. To decrease the bystander edits, we used a specific ABE (NG-ABEmax), which was delivered as mRNA. The NG-ABEmax RNA approach was validated in patient-derived rectal organoids and bronchial epithelial cells showing sufficient gene correction to recover the CFTR function. Finally, in-depth sequencing revealed high editing precision genome-wide and allele-specific correction. Here we report the development of a base editing strategy to precisely repair the 2789+5G>A mutation resulting in restoration of the CFTR function, while reducing bystander and off-target activities. Display omitted Cereseto and colleagues reported a CRISPR-Cas base editor approach to repair a splicing mutation in the CFTR gene, 2789+5G>A, which causes cystic fibrosis (CF). The genome editing tool transferred via RNA in organoids and epithelia cells from CF patients restored the CFTR activity.