In the field of remote sensing satellite image target recognition, traditional manual discernment methods and computer vision technology are low in efficiency and time-consuming. This paper improves upon the YOLOv8 network by integrating the CPCA attention mechanism and a self-developed pixel attention module, and incorporates them after the SPPF module of YOLOv8"s backbone network. The aim is to identify specific objects in remote sensing images quickly and accurately, thereby enabling the model to better adapt to different data distributions and task requirements. Experimental results show that the improved YOLOv8 model has reached an accuracy of 85.5%, recall rate of 75.9%, mAP@0.5 of 84.2%, and mAP@0.5:0.95 of 50.6%, which represents an increase of 3.2%, 1.1%, 2.6%, and 2.5% respectively compared to the original YOLOv8 model. It also has a faster detection speed when inspecting different datasets. The improved YOLOv8 network provides an effective method for enhancing the remote sensing satellite image target recognition.