Abstract: Applications of large language models (LLMs) across multiple clinical pharmacy scenarios have expanded rapidly in recent years. However, issues such as suboptimal performance on specific tasks, hallucinations, and risks related to patient privacy and data security remain prominent. As a class of methods capable of enhancing model performance, application safety, and data privacy protection from multiple dimensions, LLMs augmentation technologies are of great significance for overcoming the application bottlenecks of LLM-centered technical architectures in the field of clinical pharmacy and promoting their translation into clinical practice. This article reviewed recent studies on the application of LLMs augmentation technologies in clinical pharmacy. Among the included studies, pharmacovigilance was the most frequently investigated scenario (58.3%). Keyword cloud analysis indicated that researchers focused primarily on patients’ clinical medication use and adverse drug reactions. LLMs augmentation technologies can generally be categorized into three types: knowledge injection at the model level, knowledge augmentation during the model reasoning stage, and model system-level augmentation. Single or multiple augmentation technologies can be integrated with LLMs to construct a multidimensional augmentation framework. Existing studies have demonstrated that the augmentation technologies exert varying degrees of positive effects on improving LLM performance, optimizing clinical utility, and reducing potential safety risks. Nevertheless, at the current stage of development, LLMs augmentation technologies remain constrained by limitations in data and knowledge foundations, inadequate cross-scenario generalizability, and insufficient capabilities for complex reasoning. Furthermore, challenges persist regarding the lack of well-established evaluation frameworks and regulatory standards, as well as concerns over the safety of clinical deployment. Future research is warranted to address these critical issues and challenges, thereby further promoting the safe, standardized application and translational implementation of LLM-augmented frameworks in multi-scenario clinical pharmacy services.