<p>High utility patterns mining aims to extract significant information from transaction databases, enabling users to make more informed decisions. Meta-heuristic-based HUPM methods significantly enhance the efficiency of mining in large-scale data. The existing research reviews are mostly focused on comparing method categories, lacking summaries from key technical perspectives such as pruning strategies and population update strategies. This paper provides a review of existing mining methods, including full sets and derived efficient patterns. For full sets HUPM methods, the paper presents a pioneering summary and comparison of pruning strategies, including transaction-weighted utility upper bounds, encoding vectors, probabilistic bit vector compression, and prelarge concepts. It also comprehensively examines population update strategies from multiple perspectives, including particle swarm optimization, ant colony optimization, artificial bee colony, and artificial fish swarm methods. For derived HUPM methods, the review introduces and analyzes approaches based on top-k, closed, high-average, and fuzzy utility patterns, discussing the problems they address and their developmental status. Addressing the limitations of current meta-heuristic HUPM methods, the paper proposes future research directions, including data stream mining based on adaptive sliding windows, cross-level meta-heuristic constraints, and multimodal high utility patterns mining.</p>