This paper introduces an efficient approach to enhance object detection by utilizing the EfficientDet model and optimizing its hyperparameters using moth flame optimization (MFO). The proposed method performs fine-tuning of important parameters such as learning rate, batch size, and anchor scales to improve the accuracy of object detection over existing techniques. Integration of moth flame optimization (MFO) with EfficientDet framework allowed for dynamic hyperparameter tuning, improving the model’s detection capabilities without increasing computational costs. The proposed model is evaluated on the COCO dataset, and the results reveal that the performance of the model in detecting objects across varying scales and aspect ratios is better.