With the ongoing advancement of railway information technology, the volume of railway data is steadily increasing. This vast array of diverse data types presents significant challenges to railway data storage. In order to enhance the storage capacity of railway data, this paper proposes a secure storage architecture for railway data based on data lake. Firstly, a data classification standard is established based on business requirements and data types, and a cost-effective, high-efficiency storage scheme is proposed. Based on the scheme, a storage recommendation model suitable for railway application scenarios is constructed, and a data lake storage architecture including a source layer, an integration layer, and a market layer is designed. Then, based on business requirements and security standards, a three-tier data encryption acquisition scheme is proposed. Finally, a data access process based on data lake has been developed to ensure the secure storage of railway data.