Breast cancer is one of the most common among women. Because of its aggressive nature, it is a direct cause of women's death. Previous studies sought to detect it in different ways through thermography and mammography, but the measurements were not accurate and sufficient for this process, and some of them are very expensive, especially for developing and poor countries. Also, detection requires more than one specialized diagnosis, and this is also very rare. Therefore, this study proposes a high-resolution model that aims to detect breast cancer using thermography via a convolutional neural network algorithm. The processing images is done using computer vision techniques to standardize the colors and fix the size and gauges, then analyzed using deep learning algorithms, as they help to reduce variability and improve the reliability of the analysis. Datasets are categorized into static, dynamic, and mixed images. Three different models are proposed to handle each category. The results show that the accuracy, sensitivity, specificity, and Matthews’s correlation coefficient of static datasets are 93.1%, 83.17%, 98.1%, and 84.5%, dynamic dataset are 99.4%, 99.1% and 99.5%, and 98.7%, and mixed model are 99.9%, 99.2, 99.5% and 97.8% respectively. Through these results, the study hopes that the model will be a sign of hope to help diagnose cases early, while saving the high costs, in addition to the high accuracy of detection.