For modelling of real-time safety critical systems, when traditional testing techniques cannot be applied, formal system verification is crucial. Alvis is a modelling language that combines possibilities of formal models verification with flexibility and simplicity of practical programming languages. Solutions introduced in Alvis make the development process easier and help engineers to cope with more complex systems. The paper deals with a state-based approach to the verification of Alvis models. Until the research presented in the paper were conducted, the verification process was mostly action-based. The nuXmv tool, as one of the top model checkers, was selected for the task of state-base verification of Alvis models translated into the SMV modelling language. The paper presents a translation algorithm and usability studies performed on existing safety critical systems.