Introduction: Stroke treatment must be given within the first few hours of symptom onset, so rapid Emergency Medical Services (EMS) response is essential for positive patient outcomes. Examining EMS response characteristics across temporal and geographical perspectives is critical for improving Alabama's stroke death rate from its current 49th place in the U.S. Methods: We examined EMS delay characteristics for patients with suspected strokes from 2018-2019 across Alabama, with particular attention paid to rural and urban differences. We used EMS call data from the Alabama Department of Public Health and defined possible stroke cases as calls where a stroke scale completed by EMS was positive or indeterminate. We coded the time between EMS dispatch and destination arrival as EMS delay. This study incorporates global and local spatio-temporal weighted ordered logistic regression models to evaluate significant non-stationary correlations of factors with EMS delay by accounting for unobserved heterogeneity. Results: We examined 17,088 possible stroke cases and found that 74% of these calls had total response times within 60 min. The longest EMS delay occurred in rural counties. EMS response characteristics, such as pre-arrival instructions and advanced vehicle navigation, were associated with shorter delays, while long travel distance and on-scene/transport times were associated with longer EMS delays. The impact of response characteristics on stroke-event-based EMS delay varied significantly across rural and urban counties and time (i.e., between 2018 and 2019). Conclusion: The revealed spatio-temporal correlations are useful for EMS personnel in applying effective localized rural and urban EMS response improvement strategies.