The detection of blood cells in microscopic images is a crucial task in clinical diagnosis and hematology research. It helps identify abnormalities originating from blood conditions and other blood disorders. Different deep learning techniques are proposed to detect blood cells automatically. However, the small sizes, various shapes and other unique characteristics of blood cell images pose many research challenges. This study proposes an efficient method for blood cell detection, called ESQ-YOLO. The proposed method is an improvement of YOLOv8 which integrates Efficient Multi-scale Attention and Squeeze-and-Excitation mechanisms to help the model focus on multi-scale features and enhance the ability to learn channel weights, respectively. Furthermore, ESQ-YOLO is based on the nano version of YOLOv8 to achieve computational efficiency. The experimental results on blood cell dataset show that the proposed method outperforms other YOLO based algorithms in terms of both detection and computational performance. The source codes of ESQ-YOLO can be freely downloaded from our Github page https://github.com/letractien/ESQ-YOLO.