AbstractGiardia duodenalis is the most common intestinal protozoan in humans and animals worldwide, including eight morphologically identical assemblages, infecting pets, livestock, wildlife and human beings. Assemblages A and B are those with the higher zoonotic potential, and they have been detected in several mammals other than humans; the others (C to H) show a higher host specificity. Cats can harbour both the specific Assemblage F and the zoonotic ones A and B. Several studies have been carried out on G. duodenalis genotypes in cats; however, the role of this species in the epidemiology of giardiasis is still poorly understood. In this scenario, the present study carried out the detection and genetic characterization at sub‐assemblage level of G. duodenalis from colony stray cats in central Italy. In the period 2018–2019, 133 cat faecal samples were analysed for the presence of G. duodenalis cysts by a direct immunofluorescence assay. Positive samples were subsequently subjected to molecular analyses for assemblage/sub‐assemblage identification. Forty‐seven samples (35.3%) were positive for G. duodenalis cysts by immunofluorescence. G. duodenalis DNA was amplified at SSU‐rDNA locus from 39 isolates: 37 were positive for zoonotic Assemblage A and 2 showed a mixed infection (A + B). Positive results for the β‐giardin gene were achieved for 25 isolates. Sequence analysis revealed 16 isolates belonging to Sub‐assemblage AII and 8 to Sub‐assemblage AIII. One isolate resulted as ambiguous AI/AIII. Large sequence variability at the sub‐assemblage level was detected, with several double peaks and mutations, making complex a proper isolate allocation. When compared with previous studies, the 35.3% prevalence of G. duodenalis in cats reported in the present article was surprisingly high. Moreover, all positive cats resulted to be infected with zoonotic assemblages/sub‐assemblages, thus indicating stray cats as a possible source of human giardiasis and highlighting the sanitary relevance of cat colonies in the study area.