The main objective of this study is to determine the predictive power of mathematics self-efficacy resources and perception of science self-efficacy on academic achievement. The study, adopting a relational screening model, was conducted with a total of 698 students in sixth, seventh and eighth grade level of a state secondary school. Mathematics self-efficacy resources and science self-efficacy scales were used as data collection tools. Additionally, the half-terms report card grade point averages prior to the term when the study was conducted were also taken into account to determine students' academic achievements. Data analysis was performed by Pearson product-moment correlation technique and multiple linear regression analysis. According to the obtained results, resources of mathematics self-efficacy and perception of science self-efficacy were found to be significantly correlated with academic achievement in high, medium and low levels of influence. Belief in learning ability, belief in skills, mastery experiences, social persuasions and physiological states, which have significant impacts on the academic achievement, account for 48 % of the variance in the academic achievement. Indirect experiences, however, do not have a significant effect on academic achievement. Besides this, when examined the relationship between dimensions of science self-efficacy perception and mathematics self-efficacy resources; a significant relationship was observed between the belief in learning ability and mastery experiences, indirect experiences and physiological states, belief in skills and mastery experiences, indirect experiences and social persuasions.