Object detection is an important technology for the intelligent interpretation of remote sensing images，with extensive applications in emergency disaster response and key target monitoring.With the increase in the number of satellites and the improvement in resolution，application requirements have evolved from simple object detection to fine-grained information extraction.However，the existing object heading discrimination algorithms are primarily suitable for objects with regular contours，e.g.，ships or vehicles.They are based on the detection of an oriented bounding box，and determine object heading by classifying the directions of the four edges of the oriented box.For aircrafts in optical remote sensing images，these methods often fail due to the misalignment between the predicted bounding box edges and the aircraft’s principal axis，leading to poor angle prediction accuracy.To address this issue，we propose an object heading discrimination algorithm for aircrafts guided by an independent angle prediction branch (IAPB).Based on an improved YOLOv6 network model，our approach performs oriented object detection，and incorporates a dedicated angle prediction branch for angle estimation.Within this branch，we replace regression-based angle prediction with a classification method，effectively resolving the boundary discontinuity problem inherent in regression-based approaches.We adapt the SJTU dataset to a label format suitable for heading discrimination，and validate our method experimentally.Compared with the state-of-the-art (SOTA) model OHDet，our algorithm achieves an improvement of 3.14% in the object detection accuracy and an improvement of 6.52% in the heading discrimination accuracy.