Abstract: Time delay estimation (TDE) in impulse noise environments is one of the most active subjects in the signal processing field. In the symmetric α-stable distribution noise environments, a new time delay estimation algorithm for HF fading signal based on a nonlinear compress core function (NCCF) is provided. The NCC function is able to not only suppress the pulse stretching of the α-stable distribution noise, but also preserve the cross correlation between the object signals. As a result, it confirms the good performance of the signal time delay estimation algorithm. Comparing with the classical time delay estimation methods, this new method has the advantages in reducing the computational complexity as well as improving the stability of the estimator. Theoretical proof is also provided in this paper to illuminate the features of the algorithm. Simulation results shows that the new algorithm has better accuracy and validity than the GCC, FLOC, GCF and the nonlinear transformation algorithms.