A novel approach to developing multiplayer games using Unreal Engine 5 is proposed, focusing on efficient data transmission, minimizing latency, and synchronizing player experiences in shared game worlds. The method incorporates specialized algorithms for compensating transmission delays, network fluctuations, and packet loss, along with position prediction techniques to mitigate network latency effects. The research aims to create a cooperative adventure game as a real-time software system, demonstrating practical solutions to key challenges in data transmission and synchronization over the Internet. By adopting a listen server architecture, the approach eliminates the need for dedicated game servers (players take turns acting as servers) and reduces packet delay through localized hosting. Synchronization of game states is achieved through replication, ensuring consistent representation across clients, and Remote Procedure Calls (RPCs) for efficient data transmission. A computational experiment on local prediction of character movement was conducted to verify the quality of the prediction algorithms. Local simulation for weapons, providing players with visual effects, and physical simulation to mitigate network latency is implemented. Integration with Epic Games online services enables seamless connectivity with other players' game worlds. Extensive testing validates the effectiveness of the predictive movement system, ensuring smooth character movements even under significant network delays (500 ms) and real-world unpredictability (1 s). The proposed approach demonstrates precision and efficiency in addressing challenges in multiplayer game development, facilitated by packet retransmission mechanisms and local simulation strategies. Future work will focus on further refining prediction algorithms to accommodate various movement modes. Keywords: packet transmission latency minimization; character movement prediction; local simulation; network latency compensation; player game state synchronization.A novel approach to developing multiplayer games using Unreal Engine 5 is proposed, focusing on efficient data transmission, minimizing latency, and synchronizing player experiences in shared game worlds. The method incorporates specialized algorithms for compensating transmission delays, network fluctuations, and packet loss, along with position prediction techniques to mitigate network latency effects. The research aims to create a cooperative adventure game as a real-time software system, demonstrating practical solutions to key challenges in data transmission and synchronization over the Internet. By adopting a listen server architecture, the approach eliminates the need for dedicated game servers (players take turns acting as servers) and reduces packet delay through localized hosting. Synchronization of game states is achieved through replication, ensuring consistent representation across clients, and Remote Procedure Calls (RPCs) for efficient data transmission. A computational experiment on local prediction of character movement was conducted to verify the quality of the prediction algorithms. Local simulation for weapons, providing players with visual effects, and physical simulation to mitigate network latency is implemented. Integration with Epic Games online services enables seamless connectivity with other players' game worlds. Extensive testing validates the effectiveness of the predictive movement system, ensuring smooth character movements even under significant network delays (500 ms) and real-world unpredictability (1 s). The proposed approach demonstrates precision and efficiency in addressing challenges in multiplayer game development, facilitated by packet retransmission mechanisms and local simulation strategies. Future work will focus on further refining prediction algorithms to accommodate various movement modes. Keywords: packet transmission latency minimization; character movement prediction; local simulation; network latency compensation; player game state synchronization.