Medical Named Entity Recognition (MNER) plays a pivotal role in Natural Language Processing (NLP), particularly within the medical domain. This research presents an innovative methodology for MNER, adeptly tackling the challenges presented by the diversity of terminologies, privacy issues, and the significant costs associated with annotation in Chinese medical records. In this study, we begin by assembling a meticulously curated medical terminology database, drawing upon the expertise of domain specialists. This comprehensive database enhances the precision in comprehending medical semantic structures, enabling more accurate recognition of named entities in medical texts. To further improve the performance of MNER, we leverage the power of the pre-trained Bidirectional Encoder Representations from Transformers (BERT) language model. By fine-tuning BERT on Chinese medical diagnostic and treatment texts, we achieve proficient recognition and categorization of entities, along with their respective types. Through rigorous experimentation on a large-scale, real-world Chinese medical record datasets, we showcase the superior performance of our proposed framework. Our comprehensive and novel approach is instrumental in accurately extracting named entities from medical texts, thereby enhancing clinical decision support, organizing medical knowledge, facilitating drug discovery, and contributing to a broad spectrum of other medical applications. The results of this study demonstrate the effectiveness and potential impact of our methodology in the field of medical named entity recognition. By addressing the challenges specific to Chinese medical records, we provide valuable insights and advancements that can benefit various healthcare applications and ultimately improve patient care. This research holds significant value and relevance in the field.