Abstract: To investigate the effects of elevated temperature and cooling methods on the bond performance between steel rebars and slag–fly ash-based geopolymeric recycled aggregate concrete (SF-GRAC), 196 central pull-out specimens were tested after exposure to elevated temperatures followed by different cooling methods. The influences of temperature (20, 200, 400, and 600℃), recycled aggregate replacement ratio (0%, 30%, 70%, and 100%), anchorage length (2.5 d, 5 d, and 7.5 d), and cooling method (natural cooling and water spraying) on the bond behavior were systematically investigated. The results indicate that the bond strength initially increases and then decreases with increasing temperature, reaching its maximum at 200℃, where it is 5%–8% higher than that at room temperature. The most severe deterioration occurs at 600℃, with the residual bond strengths after natural cooling and water spraying decreasing to 61% and 42% of the corresponding room-temperature values, respectively. Compared with natural cooling, water spraying causes more pronounced degradation of the bond performance. Under the same temperature and cooling conditions, the bond strength first increases and then decreases with increasing recycled aggregate replacement ratio, and the optimum bond performance is achieved at a replacement ratio of 30%. Moreover, the bond strength decreases with increasing anchorage length. Under room-temperature and natural-cooling conditions, increasing the anchorage length from 2.5 d to 7.5 d results in a 68% reduction in bond strength, and this influence becomes less significant as the temperature increases. Scanning electron microscopy (SEM) observations reveal that elevated temperatures significantly increase the number of pores and microcracks within SF-GRAC, while water spraying further aggravates interfacial damage. Based on the experimental results, a bond–slip constitutive model considering the effects of elevated-temperature damage is established, and the predicted results show good agreement with the experimental data. The proposed model provides a basis for the application of SF-GRAC and offers theoretical support for evaluating its post-fire structural safety.