Many children have difficulties with accurate self-monitoring and effective regulation of study, and this may cause them to miss learning opportunities. In the classroom, teachers play a key role in supporting children with metacognition and learning. The present study aimed to acquire insights into how teachers' cognitive and metacognitive strategy instruction, as well as teacher-directed and child-centered instructional practices are related to children's self-monitoring accuracy, regulation of study, and learning performance. Twenty-one teachers and 308 children (2nd and 4th grade elementary school) participated. Teachers instructed a secret code task, children had to learn the match between letters of the alphabet and corresponding symbols. Teachers were observed and audio-recordings were made of their instructions. Then, children were asked to: (1) make restudy selections; (2) complete a test; and (3) self-monitor test performance. Although teachers both addressed cognitive and metacognitive strategies, they more often instructed children about cognitive strategies. Further, teaching practices were more often teacher-directed than child-centered. Although there were no relations between teachers' instructions for metacognitive strategies and children's outcome measures, teaching cognitive strategies was positively associated with children's performance and self-monitoring accuracy. However, teaching cognitive strategies did not predict effective restudy selections. Rather, child-centered instructions (i.e., giving children autonomy to regulate their own learning) positively predicted children's restudy, and further, children's self-monitoring was more accurate in classrooms where teachers more often used child-centered instructional practices. This seems to imply that not only the content of the instructions itself, but particularly the way these are given, affects children's metacognition.