Dual-port direct drive wave energy power generation systems (DP-DDWEPGS) have received widespread attention due to their smooth and zero-free output power, compared to single-port direct drive wave energy power generation systems (SP-DDWEPGS) which have the disadvantage of large out-put power fluctuations. To further enhance the performance of the DP-DDWEPGS, optimal power capture control is proposed to achieve maximum power point tracking. Meanwhile, a multiport converter is applied to the DP-DDWEPGS to solve the problem caused by an excessive number of switching devices in the overall system converter. The multiport converter fulfills all the functional requirements of the DP-DDWEPGS while reducing the number of switching devices. However, switch multiplexing of the multiport converter also introduces coupling relationships between each port and the wave force exhibits time-varying characteristics, necessitating advanced control methods with superior fast-tracking capability. Therefore, in this paper, a decoupling duty cycle optimization model predictive control for DP-DDWEPGS is proposed. Based on the characteristics of switching multiplexing, NSC finite control set model predictive control (FCS-MPC) decouples the current prediction and the cost function, reduces the number of candidate voltage vectors in each operation, and shortens the operation time by 70%. To address the issues of high ripple value and increased error due to decoupling in FCS-MPC, duty cycle optimization control is added, greatly reducing the fluctuations in electromagnetic force and power of the permanent magnet linear generator (PMLG). Based on the established simulation model, the feasibility and superiority of the multiport converter and decoupling duty cycle optimization model predictive current control method are verified.