The effective permeability impacted hydrate production severely. But its changes during hydrate production are studied few, due to no stable pressure and temperature for permeability measurement. Herein, this condition was achieved by the intermittent depressurization. The effective permeability of hydrate-bearing samples was measured. The feasibility of this method was analyzed by hydrate dissociation and saturation. Afterwards, the measured and predicted effective permeability were compared and further analyzed from hydrate distribution, preferential flow path, and fresh hydrates. The results showed that hydrate dissociation of 1.24%–2.22% did not impact measurements. Hydrate distribution, preferential flow path, and fresh hydrates were found to have a larger effect on the random changes of the effective permeability during hydrate production in comparison of hydrate saturation and pore habits. The stronger heterogeneity of hydrate distribution caused the greater randomness that the effective permeability difference was 0.346 D-0.4 D at the close hydrate saturations. Under this situation, the fixed preferential flow path was possible to form and stabilized effective permeability at 0.563 D-0.569 D. In the sample with a weaker heterogeneous hydrate distribution, the more apparent randomness was increased by fresh hydrate saturation of 1.66%–5.08%. This work may be beneficial to understand gas-water flow capacity in real settings.