Abstract: This study aims to determine the influence of salt-freezing conditions on the durability of basalt fiber fine stone concrete and accurately predict strength changes considering nonlinear characteristics and external factors. Hydraulic structures and their environmental conditions in the saline-alkali land of Jingdian irrigation area in Gansu, China served as a test case. Indoor material tests were conducted by varying the freezing and thawing medium (clean water, 3% NaCl solution, 5% Na₂SO₄ solution) and basalt fiber content (0, 0.05%, 0.10%, 0.15%, 0.20%). This provided preliminary insights into uniaxial compressive strength changes of basalt fiber fine stone concrete under different salt-freeze environments. Based on laboratory results, a basic-BP model combining BPNN and the beetle antennae search algorithm (BAS) was developed to predict compressive strength changes considering varying salt-freezing conditions. Additionally, two other BPNN models improved by intelligent algorithms were constructed for comparison. Model performance and error analysis revealed the BAS-BP model predictions agreed closely with tests, demonstrating good accuracy and stability. This can greatly improve efficiency in obtaining durability test results for basalt fiber fine stone concrete. Appropriate basalt fiber content, such as 0.15%, was found to enhance salt freezing resistance, with optimal performance across factors. NaCl exposure caused more severe damage than Na₂SO₄ during freezing and thawing. The error analysis revealed that the BAS-BP model's predictions most closely matched the test results, demonstrating strong predictive accuracy and stability.