A previously developed expert system for landscape visual quality assessment was evaluated for reliability and validity against four professional landscape architects, one of whom’s rules upon which the system was based, and two lay subjects. Results indicated that the system has good initial reliability and internal validity, but the external validity assessments appeared poor. It is believed that this was due to the subjects’ unfamiliarity with the system’s format and logical construction. For this reason, it is recommended that a training program be developed and the reliability and validity be reassessed, with and without training the subjects.