This study aimed to investigate the influence of semicircular bend (SCB) specimen size (R), loading mode (Me), and loading rate (Lr) on fracture resistance indicators, namely, fracture work (Wf), fracture energy (Gf), and fracture strength (Kf), of asphalt concrete at three different temperatures (−30°C, −20°C, and 10°C). Using Minitab software, response surface methodology (RSM) under central composite design (CCD) was employed to design experiments and develop predictive models for Wf, Gf, and Kf in terms of R, Me, and Lr at each temperature. The results demonstrated that the RSM models accurately predicted the fracture test data for all temperatures. The analysis of variance (ANOVA) revealed that R, Me, and Lr significantly influenced Wf, Gf, and Kf at each temperature, whereas the square terms R2, Me2, and Lr2 were not significant. The significance of two‐way interaction terms varied across different responses and temperatures. Overall, the experiments conducted at −30°C, −20°C, and 10°C indicated that varying R, Lr, and Me had notable effects on Wf, Gf, and Kf. Increasing R and Me while decreasing Lr resulted in an increase in Wf and Gf. Furthermore, Kf exhibited a direct relationship with R and Lr but an inverse relationship with Me.