When embarking on vocational training in the STEM field, a considerable proportion of students are required to engage with new and demanding courses. This study aimed to understand the relationship between the self-regulated learning (SRL) process and specific components of STEM fields, with a particular focus on spatial skills and technical drawing. In total, 180 students from the Swiss vocational education and training (VET) sector completed self-report instruments to assess three dimensions: motivational beliefs (self-efficacy and task value), monitoring (confidence judgment and calibration), and spatial skills. The structural equation model indicated that confidence judgments, calibration, and interest directly predicted spatial performance, while the relationship between self-efficacy beliefs and performance was mediated by confidence judgment and calibration. Moreover, the VET program duration was found to be negatively associated with task value. This study highlights the importance of considering the SRL process in the development of spatial skills.