Background. The quality of the statistics on causes of death (CoD) does not present consolidated indicators in literature further than the coding group of ill-defined conditions of the International Classification of Diseases. Our objective was to assess the territorial quality of CoD by reliability of the official mortality statistics in Spain over the years 1980-2019. Methods. A descriptive epidemiological design of four decades (1980-, 1990-, 2000-, and 2010-2019) by region (18) and sex was implemented. The CoD cases, and adjusted rates and ratios (to all-cause) were assigned by reliability to unspecific and ill-defined quality groups. The regional mortality rates were contrasted to the Spanish median by decade and sex by the Comparative Mortality Ratio (CMR) in a Bayesian perspective. Statistical significance was considered when the CMR did not contain the value 1 in the 95% credible intervals. Results. Unspecific, ill-defined, and all-cause rates by region and sex have decreased over 1980-2019. Also, the proportion of ill-defined CoD has decreased in both sexes over the decades, but was still prominent in 3 regions. CMR of ill-defined CoD in both sexes exceeded the Spanish median in 3 regions in all decades. In the last decade, the CMR in women significantly exceeded in 5 regions for ill-defined and in 6 regions for unspecific CoD, while the CMR in men exceeded in 4 of the 18 regions, respectively. Conclusions. The quality of mortality statistics of causes of death has increased over the 40 years in Spain; however, quality gaps still remain across regions. Institutions involved may consider surveilling CoD, and developing a systematic medical post-grade training on death certification.