Natural runoff reflects the effects of climate change and land surface variation, which is important for water resources allocation in the basin. Further understanding long-term variations in natural runoff is of great significance in water resources management. This study applied the empirical mode decomposition method (EMD), moving average method and anomalous accumulation method to reveal the temporal and periodical changes in natural runoff with long-term data from 1470 to 2018 in the Yellow River basin. The results showed that annual natural runoff had a significant random characteristic, showing wet-dry fluctuations within different periods. The annual natural runoff at 5% and 95% frequency were 33.70 billion m³ and 68.10 billion m³, respectively. The natural runoff of the Yellow River could be divided into two periods by the year of 1825. The EMD method showed that the natural runoff had a significant periodical cycle of 2.8 years, followed by 23.8 years, 47.3 years, 5.6 years and 11.3 years. The periodic feature was mainly influenced by the Sunspot activity, El Niño phenomenon and the amplitude changes factor of earth pole movement.