The increasing adoption of artificial intelligence (AI) in Philippine higher education has transformed teaching and learning, particularly in science education such as Electricity and Magnetism (Shi et al., 2024). However, limited empirical evidence exists on the relationship between AI tools usage and academic performance, as well as the mediating role of self-regulated learning (SRL) among pre-service science teachers (Al Darayseh, 2023; Moroianu et al., 2023). This study addressed this gap using a quantitative correlational design with mediation analysis grounded in Zimmerman’s (2002) SRL theory. A total of 66 second-year BSEd-Science pre-service teachers from a university in Eastern Visayas participated through total enumeration sampling during the second semester of A.Y. 2025–2026. Data were gathered using validated and adapted questionnaires measuring AI tools usage and SRL, while academic performance was based on final grades, where lower numerical values indicate better outcomes. Results showed that AI tools usage was “Often” (M = 3.72), SRL was “Highly Self-regulated” (M = 3.36), and academic performance was “Good” (M = 1.65). Mediation analysis revealed that AI tools usage significantly predicted SRL (β = 0.593, p < .001), indicating that greater AI engagement was associated with higher SRL. In turn, SRL significantly predicted academic performance (β = -0.436, p < .001), suggesting improved performance with higher SRL. AI tools usage also had a significant direct effect on academic performance (β = -0.279, p = .005). The indirect effect through SRL was significant (β = -0.259, p < .001), accounting for 48.2% of the total effect (β = -0.537, p < .001), indicating partial mediation. These findings highlight that AI tools enhance academic performance both directly and indirectly by strengthening SRL. These findings provide empirical evidence to guide educators and policymakers in establishing evidence-based guidelines for AI integration, ensuring future educators are equipped for technology-driven learning environments.