Pest birds represent significant threats to agricultural productivity and cause substantial economic losses around the world. Traditional pest control methods and tracking models are often laborious, harmful to the environment, and ineffective in practice. This paper explores the potential of deep learning models to improve pest bird detection and classification in agricultural settings. We evaluated the performance of three robust state-of-the-art object detection models, which are YoLOv8, Faster R-CNN, and C3Det, using various dataset configurations comprising high-resolution bird images, low-resolution bird images, and a mixed configuration. Our results show that while YOLOv8 excels in real-time detection of high-resolution bird images with a precision of 96.15%, C3Det is better detecting low-resolution bird images with a precision of 97.48%. These findings highlight the power of deep learning models to revolutionize pest-bird management.