Abstract: In response to the problems of uneven allocation of seat load resources and low efficiency of static allocation mode during peak periods of railway passenger transport, this paper studied a dynamic load balancing strategy for railway seat data based on visitation prediction. The paper constructed a visitation prediction model and combined migration cost constraints to design a dynamic load balancing strategy and automatic seat migration method, implement dynamic reallocation of seat resources. This strategy and method could effectively improve seat utilization, reduce the imbalance of load in China railway ticketing and reservation system (referred to as the ticket system), and provide new solutions for resource optimization of the ticket system.