As a critical component of nuclear reactors, the internal flow characteristics of fuel assemblies significantly impact heat transfer and safety, among which the spacer grid plays a crucial role in the flow behavior. However, due to the complexity of the spacer grid structure and the flow dynamics, traditional numerical simulations and experimental methods often require high computational and time costs, and making it challenging to analyze the flow of the working medium effectively. To address this, this work adopts two data-driven approaches, which are Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD) for reduced-order analysis of the flow field in a fuel rod bundle with spacer grid. Since the spacer grid’s mixing mainly serves to mix the fluid flow, the POD method is applied to extract the dominant modes and reduce the dimensionality of the flow field data near the mixing wing, and the DMD method is then used to analyze the temporal evolution of the flow field and identify key dynamic modes. By reducing the order of the flow field in the fuel rod bundle with spacer grid, the proposed approach effectively decreases computational costs for subsequent calculation while preserving essential flow characteristics. This approach offers valuable support for the optimal design of fuel assemblies. The results demonstrate that the method accurately captures the main structural features and dynamic behaviors of the flow field, achieving high efficiency and accuracy in dimensionality reduction.