Metacognition and emotions play a critical role in learners' ability to monitor and regulate their learning about 21st-century skills related to science, technology, engineering, and mathematics (STEM) content while using advanced learning technologies (ALTs; e.g., intelligent tutoring systems, serious games, hypermedia, augmented reality). In this article, we focus on the following: (1) presenting a succinct review of the assumptions, strengths, and weaknesses of two leading models of metacognition and emotions related to 21st-century skills typically not adopted by ALT researchers; (2) presenting and critiquing Azevedo and colleagues' extension of the information processing theory of self-regulated learning by articulating the assumptions as well as describing the advantages and disadvantages of including the macro-level, micro-level, and valence of metacognitive processes; and (3) proposing future directions and presenting implications for the design of metacognitive and affect-sensitive ALTs for 21st-century skills in STEM.