Irrigation districts consist of multi-level water conveyance networks and spatially heterogeneous farmlands, where water usage among farmlands under different canal systems exhibits game-theoretic characteristics. The pressing challenge is to develop a canal water resources optimization model that integrates dynamic channel conveyance properties with multi-objective coordination mechanisms, enabling multidimensional trade-offs among water allocation equity, conveyance loss, and main canal diversion stability. Taking a typical two-level canal system (Yonglan Sub-main Canal and its branch canals) in the Hetao Irrigation District as the research object, using the net flow rate, start time, and end time of water distribution to branch canals as decision variables, a multi-objective optimization model for canal water allocation that incorporated diversion stability, seepage loss, and water distribution equity was developed. The model advanced beyond existing canal distribution models by eliminating the assumption that the main canal must be in a “hydraulically continuous state” at water distribution onset. It incorporated main canal water transit time constraints and employed the non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) to derive optimized water allocation schemes. Results demonstrated that the model’s diversion schedule completed branch canal water distribution within 23.15 days per irrigation rotation—a 29.55% reduction compared with that of static models ignoring water transit constraints (32.86 days). This significantly shortened irrigation duration and markedly decreased transit seepage losses. Furthermore, the model achieved a water distribution equity objective value of merely 0.00227 (<1%), substantially improving upon models neglecting equity goals, thereby effectively balancing water-use benefits across farmlands served by different canals. This research established a multi-objective cooperative optimization model for canal water resources. The optimized results elucidated the trade-offs among suppressing main canal flow fluctuations, minimizing conveyance losses, and maximizing equity, providing a scientifically grounded water allocation framework for irrigation district management.