Cleidocranial Dysplasia (CCD) is a rare genetic disorder characterized by skeletal abnormalities and dental anomalies, primarily caused by variants in the RUNX2 gene. Understanding the spectrum of RUNX2 variants and their effects on CCD phenotypes is crucial for accurate diagnosis and management strategies. This systematic review aimed to comprehensively analyze the genotypic and phenotypic spectra of RUNX2 variants in CCD patients, assess their distribution across functional regions, and investigate genotype-phenotype correlations. This review included 569 reported variants and 453 CCD patients from 103 articles. Of 569 variants, in-frame variants constituted 48.68%, while null variants accounted for 51.32%. Regarding locations, RUNX2 variants were predominantly located in the RHD (55.54%), followed by PST (16.34%), NMTS (6.33%), QA (4.75%), VWRPY (1.23%), and NLS (1.41%) regions while 10.19% were in non-coding regions. In-frame variants occurred primarily in the RHD (90.97%), while null variants were found across various regions of RUNX2. Data analysis revealed a correlation between variant location and specific skeletal features in CCD patients. Missense variants, predominantly found within the functionally critical RHD, were significantly associated with supernumerary teeth, macrocephaly, metopic groove, short ribs, and hypoplastic iliac wings compared to nonsense variants. They were also significantly associated with delayed fontanelle closure, metopic synostosis, hypertelorism, limited shoulder abduction, pubic symphysis abnormalities, and hypoplastic iliac wings compared to in-frame variants found in other regions. These findings underscore the critical role of the RHD, with missense RHD variants having a more severe impact than nonsense and other in-frame variants. Additionally, in-frame insertions and deletions in RUNX2 were associated with fewer CCD features, compared to missense, frameshift, and nonsense variants. Null variants in the NLS region exhibited weaker associations with delayed fontanelle closure, supernumerary teeth, Wormian bones, and femoral head hypoplasia than variants in other regions. Moreover, the NLS variants did not consistently alter nuclear localization, questioning the role of NLS region in nuclear import. In summary, this comprehensive review significantly advances our understanding of CCD, facilitating improved phenotype-genotype correlations, enhanced clinical management, and a deeper insight into RUNX2 functional domains. This knowledge has the potential to guide the development of novel therapeutic targets for skeletal disorders.