With the development of infrastructure, there are growing numbers of high geothermal environments, which, therefore, form a serious threat to tunnel structures. However, research on the changes in mechanical properties of shotcrete under high temperatures and humid environments are insufficient. In this paper, the combination of various temperatures (20 °C/40 °C/60 °C) and 55% relative humidity is used to simulate the effect of environment on the strength and stress–strain curve of basalt fiber reinforced shotcrete. Moreover, a constitutive model of shotcrete considering the effect of fiber content and temperature is established. The results show that the early mechanical properties of BFRS are improved with the increase in curing temperature, while the compressive strength at a later age decreases slightly. The 1-day and 7-day compressive strength of shotcrete at 40 °C and 60 °C increased by 10.5%, 41.1% and 24.1%, 66.8%, respectively. The addition of basalt fiber can reduce the loss of later strength, especially for flexural strength, with a increase rate of 11.9% to 39.5%. In addition, the brittleness of shotcrete increases during high temperature curing, so more transverse cracks are observed in the failure mode, and the peak stress and peak strain decrease. The addition of basalt fiber can improve the ductility and plasticity of shotcrete and increase the peak strain of shotcrete. The constitutive model is in good agreement with the experimental results.

Climate-change-induced extreme environments exacerbate pavement degradation in arid regions, where traditional concrete incurs 23~40% higher life-cycle costs due to premature cracking. Particularly in the Gobi Desert, concrete pavements suffer from conflicting performance requirements—high flexural-to-compressive strength ratio (R_f/R_c), low shrinkage, and controlled porosity—with traditional design methods failing to address multi-objective trade-offs. Existing optimization methods have proven insufficient for such complex environments, with conventional approaches addressing only individual parameters or employing subjective weighting techniques that fail to capture the interrelated nature of critical performance indicators. This study develops an integrated optimization framework combining Response Surface Methodology (RSM), Non-dominated Sorting Genetic Algorithm III (NSGA-III), Criteria Importance Through Intercriteria Correlation (CRITIC) weighting, and VIšekriterijumsko KOmpromisno Rangiranje (VIKOR) decision-making to optimize the mix proportions water–cement ratio (W/C), sand ratio, and an air-entraining agent (AEA) for sustainable pavement concrete. Response Surface Methodology (RSM) analysis via Box–Behnken design revealed distinct parameter dominance: AEA exhibited the strongest non-linear effects on R_f/R_c and porosity, while W/C primarily governed shrinkage. NSGA-III generated 73 Pareto-optimal solutions, with CRITIC selecting an optimal mix (W/C = 0.35), sand ratio = 36%, AEA = 0.200%) validated experimentally (R_f/R_c = 0.141), shrinkage = 0.0446%, porosity = 2.82%. Microstructural characterization using scanning electron microscopy and low-field nuclear magnetic resonance (SEM/LF-NMR) demonstrated refined pore distribution and enhanced compactness. This framework effectively resolves trade-offs between performance indicators, providing a scientifically robust method for designing durable pavement concrete that reduces shrinkage by 13.0% and porosity by 13.5% compared to conventional mixes, lowering maintenance costs in arid regions.