Objective: To analyze the spatial patterns of distribution of Specialized Rehabilitation Centers(CERs). Methods: An ecological-level study design, descriptive and Bayesian spatial regression analyses was carried out using the 438 Brazilian Health Regions. Main outcome is the “presence or absence of CER” and as covariates the rate of PWD, health funding and socioeconomic indicators. Modeling considered a binomial probability distribution and spatial randomness in the Besag-York-Mollié model. Neighborhood relationships between Health Regions were constructed using Queen-type contiguity matrix. Results:Considering four types of disability(hearing, intellectual, physical and visual) alone, the rate of PWD and transfers of financial resources increase the chances of implementing an CER. In joint analysis of disabilities, implementation of CERs doesn’t correspond to regions with higher rates of PWD. Conclusions:Prioritize the CER implementation according to territorial distribution of needs is essential to improve equitable access. Use of qualified data to identify PWD is essential for the creation of public policies.