Considering how congestion will propagate in the near future, understanding traffic congestion propagation has become crucial in GPS navigation systems for providing users with a more accurate estimated time of arrival (ETA). However, providing the exact ETA during congestion is a challenge owing to the complex propagation process between roads and high uncertainty regarding the future behavior of the process. Recent studies have focused on finding frequent congestion propagation patterns and determining the propagation probabilities. By contrast, this study proposes a novel time delay estimation method for traffic congestion propagation between roads using lag-specific transfer entropy (TE). Nonlinear normalization with a sliding window is used to effectively reveal the causal relationship between the source and target time series in calculating the TE. Moreover, Markov bootstrap techniques were adopted to quantify the uncertainty in the time delay estimator. To the best of our knowledge, the time delay estimation method presented in this article is the first to determine the time delay between roads for any congestion propagation pattern. The proposed method was validated using simulated data as well as real user trajectory data obtained from a major GPS navigation system applied in South Korea.