Abstract: Maritime equipment group gaming (MEGG) serves as a pivotal framework for exploring and investigating the gaming strategy interaction behaviors among various groups of maritime equipment operating in complex oceanic task scenarios, such as adversarial simulation, maritime traffic, and maritime rescue. The equipment groups typically include manned and unmanned vessels, carrier-based aircraft, and other similar assets, and the reviewed gaming strategy interaction is primarily concerned with confrontation, competition, and cooperation. First, the survey clarifies the conceptual boundaries of maritime equipment group gaming, including the differences between theories such as population game, swarm intelligence, and other related paradigms. This conceptual clarification helps lay the groundwork for subsequent classification and analysis. Next, the survey reviews the typical task scenarios of MEGG approaches, including adversarial deduction, maritime traffic, and maritime rescue. Each of these scenarios reflects distinct operational challenges and objectives within the MEGG context. Following this, the survey further distinguishes the basic categories of MEGG approaches in detail. Specifically, MEGG approaches are categorized along multiple dimensions: game modes (including confrontation gaming, competition gaming, cooperation gaming, and mixed gaming), game scopes (covering intra-group, inter-group, and dual-level gaming), equipment heterogeneity (encompassing both single-class and cross-class systems), and intelligence levels (ranging from non-intelligent gaming to intelligent gaming). Second, the survey dissects technological progress in two main representative aspects of non-intelligent MEGG and intelligent MEGG. For non-intelligent MEGG, the survey reviews the classical methods that include Lanchester's laws, population game theory, and crowd simulation models. For intelligent MEGG, the survey reviews the MEGG approaches from aspects of traditional machine learning (covering areas such as decision support, scheduling, planning, deduction, prediction, and so on) and multi-agent reinforcement learning (focusing on adversarial deduction, task planning, and so on). At the end, the survey summarizes the current challenges in MEGG research areas and proposes six promising research directions: human-machine integrated intelligent gaming decision-making framework, trustworthiness and interpretability of intelligent gaming models, deep reasoning for maritime missions via large-scale models, hierarchical collaborative gaming mechanism for intelligent agents; standardized management system for heterogeneous intelligent agent clusters, and high-fidelity gaming systems enhanced by cross-domain expert knowledge. These six directions collectively provide structured, actionable guidance for future studies in this emerging and important field.