This study proposes a deep learning (DL)-based approach to detect malaria parasites and leukocytes in thick blood smears to support malaria diagnosis. The method uses data augmentation to increase the size of the dataset and employs the YOLOv8 algorithm for model training and parasite counting. The model achieved 95% accuracy in detecting parasites and 98% accuracy in detecting leukocytes. The results showed that the model can detect parasites and leukocytes with high accuracy, and the time required for the model to report parasitemia is significantly less than that of malaria experts. The system would be supportive for areas with poor access to healthcare. The model was compared with malaria experts, showing high agreement between the results. The model's performance was evaluated using metrics such as accuracy, sensitivity, specificity, mAP, and R². The results demonstrated that the model's parasite density estimates were in high agreement with those of experts (93% agreement). The model also showed faster processing times compared to experts, with an average of 29.88 seconds per 50-image packet, significantly faster than the experts' average of 518.33 seconds. The study highlights the potential of AI in improving the accuracy and efficiency of malaria diagnosis. The model was trained on a dataset of 333 thick blood smear images and used data augmentation to generate 666 images for training. The model's performance was evaluated using both quantitative and qualitative metrics, showing that it outperformed previous studies in terms of accuracy and sensitivity. The study also discusses the limitations of the model, including the potential for overestimation of parasite density due to cell clumps or staining artifacts. The results suggest that the model could be a valuable tool in improving malaria diagnosis and treatment.This study proposes a deep learning (DL)-based approach to detect malaria parasites and leukocytes in thick blood smears to support malaria diagnosis. The method uses data augmentation to increase the size of the dataset and employs the YOLOv8 algorithm for model training and parasite counting. The model achieved 95% accuracy in detecting parasites and 98% accuracy in detecting leukocytes. The results showed that the model can detect parasites and leukocytes with high accuracy, and the time required for the model to report parasitemia is significantly less than that of malaria experts. The system would be supportive for areas with poor access to healthcare. The model was compared with malaria experts, showing high agreement between the results. The model's performance was evaluated using metrics such as accuracy, sensitivity, specificity, mAP, and R². The results demonstrated that the model's parasite density estimates were in high agreement with those of experts (93% agreement). The model also showed faster processing times compared to experts, with an average of 29.88 seconds per 50-image packet, significantly faster than the experts' average of 518.33 seconds. The study highlights the potential of AI in improving the accuracy and efficiency of malaria diagnosis. The model was trained on a dataset of 333 thick blood smear images and used data augmentation to generate 666 images for training. The model's performance was evaluated using both quantitative and qualitative metrics, showing that it outperformed previous studies in terms of accuracy and sensitivity. The study also discusses the limitations of the model, including the potential for overestimation of parasite density due to cell clumps or staining artifacts. The results suggest that the model could be a valuable tool in improving malaria diagnosis and treatment.