Agricultural human-computer interaction, Multimodal interaction, Large language models, Smart agriculture, Knowledge graph, Human-centered AI, Hybrid AI architecture Abstract: As agriculture embraces digital transformation, the demand for intelligent, user-friendly Human-Computer Interaction (HCI) technologies is growing. This review introduces a novel three-layer integrated framework for agricultural HCI information services. The framework comprises semantic environments, multimodal interaction mechanisms, and intelligent service applications. Unlike prior reviews that address isolated modules, this work uniquely synthesizes advances in knowledge graphs, voice-text and image-based interaction, and large language models (LLMs) into a coherent system. Key technological transitions— from rule-based systems to hybrid AI models—are critically evaluated using real-world cases, including pest diagnosis, irrigation decisionmaking, and breeding optimization. To overcome persistent challenges such as fragmented data, domain adaptation limitations of general LLMs, and high interaction thresholds for farmers, we propose a hybrid AI architecture combining LLMs and domain-specific small models, enhanced by federated learning and low-code multimodal interfaces. Furthermore, this review outlines future directions, including the development of AIpowered agricultural digital humans and context-aware intelligent agents, aiming to reduce user cognitive load and enhance accessibility. By https://mc.manuscriptcentral.com/jcmse Journal of Computational Methods in Science and Engineering For Peer Review establishing this unified architecture and providing a critical synthesis of technological transitions, this study offers both a conceptual framework and a practical roadmap for next-generation Agricultural Intelligence-asa-Service (AIaaS) platforms, distinguishing itself from prior modular reviews. Page 1 of 22 https://mc.manuscriptcentral.com/jcmse Journal of Computational Methods in Science and Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For Peer Review Agricultural Human-Computer Interaction Information Service Technologies Evolution, Challenges, and Prospects Qingfeng Wei 1, Changshou Luo1,* Rupeng Luan1,* Feng Yu1, Ruifang Zhao1, Jun Y1, Yan Lu1, Furong Wang1 Beijing Academy of Agricultural and Forestry Sciences Institute of Data Science and Agricultural Economics , Beijing, 100097, China Email：weiqingfeng201@163.com; luochangshou@163.com; luanrupeng@163.com; yufeng@163.com; zhaoruifang@163.com; yujun@163.com; luyang@163.com; wangfurong@163.com Abstract: As agriculture embraces digital transformation, the demand for intelligent, user-friendly HumanComputer Interaction (HCI) technologies is growing. This review introduces a novel three-layer integrated framework for agricultural HCI information services. The framework comprises semantic environments, multimodal interaction mechanisms, and intelligent service applications. Unlike prior reviews that address isolated modules, this work uniquely synthesizes advances in knowledge graphs, voice-text and image-based interaction, and large language models (LLMs) into a coherent system. Key technological transitions—from rule-based systems to hybrid AI models—are critically evaluated using real-world cases, including pest diagnosis, irrigation decision-making, and breeding optimization. To overcome persistent challenges such as fragmented data, domain adaptation limitations of general LLMs, and high interaction thresholds for farmers, we propose a hybrid AI architecture combining LLMs and domain-specific small models, enhanced by federated learning and low-code multimodal interfaces. Furthermore, this review outlines future directions, including the development of AI-powered agricultural digital humans and context-aware intelligent agents, aiming to reduce user cognitive load and enhance accessibility. By establishing this unified architecture and providing a critical synthesis of technological transitions, this study offers both a conceptual framework and a practical roadmap for next-generation Agricultural Intelligence-as-a-Service (AIaaS) platforms, distinguishing itself from prior modular reviews. Keywords: Agricultural human-computer interaction; Multimodal interaction; Large language models; Smart agriculture; Knowledge graph; Human-centered AI；Hybrid AI architecture 1.