Since CRISPR‐based genome editing technology works effectively in the diploid frog Xenopus tropicalis, a growing number of studies have successfully modeled human genetic diseases in this species. However, most of their targets were limited to non‐syndromic diseases that exhibit abnormalities in a small fraction of tissues or organs in the body. This is likely because of the complexity of interpreting the phenotypic variations resulting from somatic mosaic mutations generated in the founder animals (crispants). In this study, we attempted to model the syndromic disease, campomelic dysplasia (CD), by generating sox9 crispants in X. tropicalis. The resulting crispants failed to form neural crest cells at neurula stages and exhibited various combinations of jaw, gill, ear, heart, and gut defects at tadpole stages, recapitulating part of the syndromic phenotype of CD patients. Genotyping of the crispants with a variety of allelic series of mutations suggested that the heart and gut defects depend primarily on frame‐shift mutations expected to be null, whereas the jaw, gill, and ear defects could be induced not only by such mutations but also by in‐frame deletion mutations expected to delete part of the jawed vertebrate‐specific domain from the encoded Sox9 protein. These results demonstrate that Xenopus crispants are useful for investigating the phenotype–genotype relationships behind syndromic diseases and examining the tissue‐specific role of each functional domain within a single protein, providing novel insights into vertebrate jaw evolution.This article is protected by copyright. All rights reserved.