The paper presents CNTCB-YOLOv7, an advanced forest fire detection model that integrates ConvNeXtV2 and CBAM to enhance the performance of the YOLOv7 algorithm. The study aims to address the challenges of detecting subtle features and complex backgrounds in large-scale forest fire areas. By incorporating ConvNeXtV2 and Conv2Former into the YOLOv7 backbone, the model's feature representation and detection accuracy are significantly improved. Additionally, the ELAN-CBAM structure is introduced to enhance the network's ability to focus on crucial information and minimize background interference. Experimental results show that the CNTCB-YOLOv7 algorithm outperforms YOLOv7 in terms of accuracy (86.18%), recall rate (0.73%), and average precision (1.14%). The model also reduces parameter count by 3.47 million, improving inference speed. The proposed method provides valuable support for forest fire behavior research and management, enabling more proactive and targeted firefighting strategies. Future work will focus on further refining the model, conducting comprehensive comparisons, and expanding evaluation criteria to enhance its generalizability and applicability.The paper presents CNTCB-YOLOv7, an advanced forest fire detection model that integrates ConvNeXtV2 and CBAM to enhance the performance of the YOLOv7 algorithm. The study aims to address the challenges of detecting subtle features and complex backgrounds in large-scale forest fire areas. By incorporating ConvNeXtV2 and Conv2Former into the YOLOv7 backbone, the model's feature representation and detection accuracy are significantly improved. Additionally, the ELAN-CBAM structure is introduced to enhance the network's ability to focus on crucial information and minimize background interference. Experimental results show that the CNTCB-YOLOv7 algorithm outperforms YOLOv7 in terms of accuracy (86.18%), recall rate (0.73%), and average precision (1.14%). The model also reduces parameter count by 3.47 million, improving inference speed. The proposed method provides valuable support for forest fire behavior research and management, enabling more proactive and targeted firefighting strategies. Future work will focus on further refining the model, conducting comprehensive comparisons, and expanding evaluation criteria to enhance its generalizability and applicability.