Whereas userÂ’s requirements for data of quality are constantly increasing, spatiotemporal databases often represent data coming from semantically heterogeneous and evolving data sources, causing answers to spatiotemporal queries to be imprecise, contradictory or even impossible to obtain because of the lack of semantic link between the different databases created at different time. Because quality of data may be crucial in a decision-making context, quality of answers to spatiotemporal queries must be considered. In this paper, we propose a semantic model of mapping that incorporate quality of data to automatically discover and establish semantic relationships between the different databases, in order to improve quality of answers to spatiotemporal queries. As a second contribution, we then propose a quality criterion adapted for the evolution of specifications of spatiotemporal databases. Finally, in order to assess the quality of the proposed model, we conducted a comparative study between the proposed model and other semantic models, consisting in a controlled evolution simulation. In the last section of this paper, we present results and conclusions drawn from these tests.