Objective: The vertical axis wind turbine (VAWT) cluster is a floating structure that integrates multiple VAWTs onto a single foundation. This design aims to maximize marine space utilization and lower the cost of floating wind power generation. The internal layout of the wind turbines and the pitching motion of the floating platform can both significantly impact the power generation efficiency of the turbines. To examine these combined effects, this study constructs three-dimensional computational fluid dynamics models for both a single wind turbine and a VAWT cluster. Methods: These models use an H-type straight-bladed turbine. The pitching motion follows a simple harmonic motion pattern with a constant period and amplitude.The computational domain for the VAWT cluster includes a stationary rectangular background domain and two movable cylindrical rotating domains. A cut-cell mesh is applied to the background domain, while polyhedral meshes are used in the rotating domains to reduce discretization errors and enhance computational accuracy. Additionally, dense prismatic boundary layer meshes are generated near the turbine blades, and local mesh refinement is introduced around the turbine wakes and rotor areas to capture flow details more accurately. Data exchange between the stationary and rotating domains is achieved using overlapping mesh techniques. The independence of the mesh is verified by evaluating the simulated instantaneous power coefficients of three mesh configurations with varying grid densities. The power coefficient for a single wind turbine is calculated across different tip-speed ratios, and the numerical modeling method is validated by comparing these results with experimental data. Subsequently, a second wind turbine is introduced into the fluid domain to form a cluster, and periodic pitching motion is applied to the cluster base. Two typical layouts (side-by-side and staggered) of the dual-rotor floating VAWT cluster are analyzed under pitching motion. Flow field characteristics and power variation patterns are examined under three rotation modes for each layout. Results: The side-by-side layout demonstrates a 5.64% maximum increase in average power coefficient during the pitching period, while the staggered layout achieves a 6.69% increase. The increases in the average power coefficient during the pitching period for the three rotation modes are 4.04%, 5.21%, and 5.64%, respectively. However, in the staggered layout, the downstream turbine is affected by the tip vortices of the upstream turbine, and the rotation mode has a more significant impact. The corresponding power coefficient increases are 5.35%, 6.69%, and 3.36%, respectively. Conclusions: Both side-by-side and staggered layouts deliver better power performance compared to a single fixed wind turbine. The rotation mode of the turbines weakly affects the average power coefficient of the side-by-side dual-rotor wind turbine cluster. Pitching motion significantly affects turbine rotor torque at both peak and valley positions. When the turbine swings into the wind, the peak positive torque increases significantly. Conversely, when it swings downwind, the peak positive torque decreases, but the valley negative torque also decreases. The overall efficiency of positive and negative torques within a pitching cycle is higher than that of a single fixed wind turbine.