The aim of this study was to adapt and validate a Spanish version of the Maryland Assessment of Recovery Scale (MARS-12). It was carried out in strict accordance with internationally recognized guidelines for test adaptation. A preliminary Spanish version of the MARS-12 was first produced through a standardized translation/back-translation process, ensuring semantic, linguistic, and contextual equivalence with respect to the original scale. Its psychometric properties were then examined in a sample of 325 people with serious mental illness recruited from six different provinces in the Basque Country (northern Spain) and Catalonia (north-eastern Spain). They were users of a total of 20 community rehabilitation and psychiatry services. Confirmatory factor analysis supported a unidimensional structure, consistent with the original scale. Scores on the MARS-12 were positively correlated (.83) with scores on the Questionnaire about the Process of Recovery, supporting convergent validity, while validity evidence based on relationships with other variables was provided by positive correlations between MARS-12 scores and scores on the Dispositional Hope Scale (.82) and on the three dimensions of the Multidimensional Scale of Perceived Social Support (range .30 to .41). Reliability of MARS-12 scores was high (McDonald's ω = .97), as was temporal stability across a one-week interval (.89). The Spanish version of the MARS-12 is a valid and reliable scale that may be used by mental health professionals to assess recovery among Spanish people with serious mental illness.