The Named Entity Recognition (NER) task is an important foundational task for information extraction in Natural Language Processing (NLP). In the field of cyber security threat Intelligence Analysis, cyber security NER is also a crucial foundational task. The evolution of methods in cyber security NER closely follows the development of NLP techniques. After the emergence of pre-trained language models, they were rapidly applied to NER tasks, demonstrating overall superior performance compared to other former methods. Since most pre-trained language models are trained on general text, researchers have attempted to supplement domain knowledge in the application of pre-trained language models. This paper proposes a Cyber Security NER model based on BERT feature-based approach with different downstream neural networks, supplemented with augmented knowledge in cyber security field. Experiments on two datasets indicate that this model can effectively leverage augmented knowledge on the NER task in different types of downstream neural networks.