Background/Objectives: Kazakhstan experienced one of the largest measles epidemics in the WHO European Region in 2023–2024 and a pertussis resurgence in 2024, while documented parental vaccine refusals kept accumulating. Subnational, register-based evidence on refusal is lacking for Central Asia. We assessed trends, geography, structure and data quality of registered refusals and their ecological links with coverage and disease incidence. Methods: Nationwide ecological study of the official refusal register (20 regions; stocks on 1 January 2024–2026), surveillance reports and population data. We estimated prevalence per 100,000 with exact Poisson 95% confidence intervals (CIs) and rate ratios (RRs), and applied χ² tests, Ward clustering, Moran’s I with local indicators of spatial association, Spearman correlation, negative binomial regression and Bland–Altman analysis. Results: Registered refusals rose from 39,792 to 66,240 (198.6 to 323.2 per 100,000; RR 1.63, 95% CI 1.61–1.65). Prevalence varied 23-fold between regions (39.8 in Turkestan to 932.4 in Atyrau) and increased significantly in 16 of 20 regions (Kostanay, RR 2.79). Personal beliefs dominated (62.1%); religious refusals (15.5%) clustered in the west and south (Moran’s I = 0.39, p = 0.011); media-attributed refusals grew by 39.8%. In 2024, refusal prevalence was associated with measles incidence (ρ = 0.55; incidence rate ratio 1.18 per 1/1000, 95% CI 1.03–1.34), and refusal explained 54.7% of pertussis cases. The register recorded 3.35–3.84 times more bacille Calmette–Guérin (BCG) refusals than routine coverage reports, and 46 of 180 cumulative vaccine-specific counts decreased. Conclusions: Vaccine refusal in Kazakhstan is growing, regionally heterogeneous in magnitude and motive, and tracked measles burden during the epidemic. Reason-tailored regional strategies and a validated, individual-level electronic refusal register are needed.