The relatively high number of students in the STEM fields not completing their program successfully makes the investigation of students’ study success particularly important in these disciplines. The present study analyses predictors of student drop-out in different STEM study programs. Subject-related prior knowledge, mathematical knowledge, domain-specific interest, study motivation (expectancy), student engagement, and satisfaction with study content are identified as relevant predictors of both drop-out intentions as well as actual drop-out risk. Satisfaction with study content proves to partially mediate the effects of the other predictors on drop-out intentions. In turn, drop-out intentions represent an early indicator of later actual drop-out risk. Results show similarities as well as differences between study programs regarding the significance of the predictors.