Abstract: Household items detection is the key of home service robots to perform target grasping tasks. In order to solve the problem of difficult detection due to the complex background of daily household scenes, dense household items, and many types of small target, we propose a method of household objects detection based on the fusion of multi-sensory mechanism. Firstly, the coordinate attention mechanism module is added to the neck part of YOLOX to reduce the influence of complicated information. Secondly, the dynamic head based on multi-sensory fusion is used to improve the detection head of YOLOX to enhance the detection accuracy of small target items. Finally, focal loss is added to the loss function to reduce the error caused by the imbalance of positive and negative samples. In PyTorch environment, the improved YOLOX algorithm was used to perform ablation experiments on a self-made household items dataset and compared with six other object detection algorithms. The experimental results show that the proposed method achieves 58.34% mAP and frame rate of 45.35 frames per second for the detection of household items, which effectively improves the detection accuracy of household items while ensuring the real-time detection of the algorithm.