In recent decades, China has experienced successive epidemics of seasonal Japanese encephalitis (JE), with the Japanese encephalitis virus (JEV) particularly spreading continuously in rural and suburban areas.Nationwide data on 9061 JE cases, mosquito abundance from 89 surveillance sites, and population movement between 337 cities during 2013-19 were obtained. Seasonal multivariate linear regression models including time trends and reconciliation terms representing annual and semiannual cycles were fitted to the weekly time series of JE cases, and the amplitude and peak time of the cycles were estimated. A metapopulation network model of inter-city population mobility coupled with an iterative Bayesian inference algorithm was established to simulate the epidemic dynamics of JEV and estimate the time-varying transmission parameters.The timing of the annual peak of JEV epidemics varied with latitude (p-value < 0.05), mainly characterized by earlier in southern cities and later in northern cities. There was no significant difference in the annual amplitude fluctuations of JEV epidemics in different latitudes (p-value > 0.05). Regions with higher values of effective reproduction number Reff were mainly concentrated in central China, including Sichuan, Chongqing and Shaanxi provinces, with the annual activity peak typically occurring around August. Infections caused by population mobility mainly occurred in hub cities with high connectivity and radiated to surrounding cities.Findings from this nationwide study can help enhance situational awareness of the spread of JE and inform appropriate intervention strategies to advance the goal of JE elimination.Copyright © 2025 The Author(s). Published by Elsevier Ltd.. All rights reserved.