Valley Fever, a respiratory disease caused by the Coccidioides fungus, presents a persistent public health challenge in arid regions like the Southwestern United States. Its prevalence is strongly influenced by environmental conditions, particularly those facilitating fungal spore dispersion, such as windborne dust. To address the critical need for advanced detection and prevention mechanisms of Valley Fever, this study leverages the capabilities of Convolutional Neural Networks (CNNs) to identify hazardous dust conditions associated with Coccidioides dispersal using NASA’s Earth Observing Data and Information System (EOSDIS) Worldview satellite imagery. The results indicate that the CNN outperforms traditional approaches, such as those utilizing Aerosol Optical Depth (AOD) and PM10 measurements, by achieving an 88\% accuracy in near real-time dust event detection. This capability supports developing early warning systems to mitigate Valley Fever risks and improve public health responses. This research bridges computer science and epidemiology to create a novel tool for predicting Valley Fever outbreaks and enabling timely interventions.