Serious games are generally considered to induce positive effects in the areas of learning motivation and learning gains. Yet few studies have examined how these factors are related specially in the context of visual aesthetic studies. Therefore, an empirical study was conducted to test the relationship between anticipated enjoyment and willingness to play, as well as between game enjoyment, self-reported cognitive and motivational learning gains and test results. In an explorative study, 25 students from undergraduate course in architecture school played the learning game "Temple Lego." The results of pre-test and post-tests were analyzed using multiple linear regressions. The analysis showed that anticipated enjoyment played only a minor part in students' willingness to learn with serious games. Of greater importance was the students' expectation that the learning game would be easy and instructive. There was a correlation between enjoyment and the post-test scores to continue being engaged with the subject matter of the game. We discuss the findings and pedagogical implications that can be applied to developing serious virtual reality (VR) games to learn about similar historical structures and contexts which is novel and sparsely explored.