To understand the impact of two learner characteristics—metacognition and goal orientation—on problem-solving, this study investigated 159 undergraduate learners’ metacognition, goal orientations, and problem-solving performances and processes in a laboratory setting using a Serious Game (SG) environment—Alien Rescue (AR)—that adopts Problem-based Learning (PBL) pedagogy for teaching space science. Utilizing multiple data sources, including computer log data and problem-solving solution scores within the SG, survey data, gameplay screencast videos, and interview data, this study combined a sequential mixed method design and serious games analytics techniques to answer the following two questions: (a) To what extent are learner problem-solving performance differences based on learner characteristics, and why? (b) To what extent are learner problem-solving process differences based on learner characteristics, and why? The results indicated that (a) learner metacognition affected problem-solving. Specifically, there were statistically significant differences in learner problem-solving performances based on metacognition, and learners also demonstrated different problem-solving processes based on metacognition. (b) Learner goal orientation impacted problem-solving. Particularly, learners in different goal orientation groups had different problem-solving processes. (c) The interaction between metacognition and goal orientations had an impact on learner problem-solving performances. Specifically, learners were clustered into three groups based on these two characteristics, including (a) high metacognition and high multiple goal orientations, (b) low metacognition and medium multiple goal orientations, and (c) medium metacognition and low multiple goal orientations. Learner problem-solving performances were statistically significant based on these three clusters. In addition, learner metacognition and goal orientations together could predict learner problem-solving performances. (d) The interaction between metacognition and goal orientations also had an impact on learner problem-solving processes. These differences in learner problem-solving performances and processes can be explained by learner characteristic differences, the problem complexity, SG design, and Dunning-Kruger effects (i.e., the cognitive bias that people of low metacognitive ability might mistakenly assess their metacognitive level as higher than it is). In addition, this study summarized 10 steps of how to be a successful and efficient problem solver in AR. These steps are as follows: 1) identify the problem correctly; 2) explore the 3D environment by visiting all rooms in AR and look over all tools; 3) discover what one alien species needs to survive in Alien Database; 4) search the Solar System Database for possible planets; 5) develop hypotheses about where this alien species can live; 6) figure out if there is any missing information needed for making a decision; 7) launch probes to gather information in the Probe Design room; 8) check the data from the probe in the Mission Control room; 9) decide whether the selected planet is a good choice for the selected alien species; 10) if so, write a recommendation message with the justification in the Communication Center—if not, go back to step 4. This research offers additional understanding of learner characteristic impacts on problem-solving in SG environments with PBL pedagogy. It can also contribute to future designs of these environments to benefit learners based on their metacognitive levels. In addition, the study limitations and further research in this area are discussed.