<p>Transporting hydrogen through in-service natural gas pipelines is the most effective method for achieving large-scale and long-distance hydrogen transportation. For the calculation method of the remaining strength of in-service natural gas pipelines with defects under hydrogen-blended conditions, the variation laws of mechanical properties of X70 material under different hydrogen enrichment ratios were analyzed. A numerical calculation model for defective pipelines was established, and the relationships between hydrogen content, defect geometric parameters, and the remaining strength of in-service natural gas pipelines were studied. Based on the dataset of the GA-BP neural network and numerical calculation results, a prediction model for the residual strength of defective pipelines under hydrogen blending conditions was established. The research results show that with the increase of hydrogen content, both the yield strength and tensile strength of X70 material exhibit a downward trend. Under hydrogen-doped environments of 20% and 50%, the yield strength decreased by 3.0211% and 6.9486%, respectively, and the tensile strength decreased by 1.3812% and 6.6298%, respectively. The residual strength of defective pipelines is negatively correlated with the defect depth, defect length, and hydrogen content. The influence of the arc length of the defect circumference on the residual strength of the pipeline is relatively small. The established prediction model can predict the residual strength of pipelines under different conditions quite well, and the maximum error rate of the prediction results is only 3.8849%. This research can provide reference value for the failure analysis and prediction of hydrogen-blended natural gas pipelines.</p>