<p>Although the durability of concrete structures has been extensively studied, key challenges remain, including the inability to account for time-dependent resistance degradation, computational complexity, and limited practicality in engineering design. To address these issues, this study reconstructs the limit state equation using support vector machine regression, leveraging its high-dimensional processing capability and small-sample adaptability to enhance current concrete durability design methodology. And then a durability design method accounting for time-dependent resistance deterioration was developed through the implementation of importance sampling, where response surface equations were utilized as surrogate limitstate functions. To facilitate engineering applications, the method proposed in this study was programmed using Matlab and equipped with a user-friendly interface developed using VB, thereby visualizing this concrete durability design methodology. Post-operation GPR detection of durability parameters facilitates the update of time-varying model parameters. Iteration of the model then enables adaptive adjustment of the durability design.The proposed concrete durability design method was validated through practical engineering applications, demonstrating reliable computational performance and user-friendly visualization procedures. The findings of this study provide theoretical foundations for concrete durability design while demonstrating significant practical applicability in engineering contexts.</p>