Stroke is a major cause of death and adult neurologic disability worldwide. Underdiagnosing a stroke at emergency may have devastating consequences on patients. We aim to develop a diagnosis score for stroke at emergency among the Lebanese population by using stroke risk factors and symptoms. A hospital-based case-control study was conducted for the score generation. Data were collected through a designed data collection sheet at two Lebanese tertiary hospitals between January 1st, 2012 and December 31st, 2014. A logistic regression determined the association of stroke risk factors and symptoms, and the rounded coefficients generated the Diagnosis Score for Stroke (DS-Stroke). Another case-control study was conducted for the score validation. Data were collected through a standardized questionnaire at five different Lebanese tertiary hospitals between January 1st, 2015 and December 31st, 2016. DS-Stroke was validated by comparing it to hospitals final diagnosis. DS-Stroke was constructed with good properties, comprising 7 items. The area under the curve was high at 0.980 (0.969–0.990; p<.001). A score <4 points indicated a high negative predictive value of stroke of 97.3%. A score ≥4 points indicated a positive predictive value of stroke of 91.3%. In the validation study, DS-Stroke had an area under the curve of 0.964 (95% CI, 0.951–0.978; p<.001), 93.2% sensitivity, 91.9% specificity, and 84.1% positive predictive value and 96.7% negative predictive value. In conclusion, the DS-Stroke is a useful tool for emergency physicians, non-specialists and front-line clinicians. It is a good diagnosis tool for use to diagnose stroke at emergency, call for further emergency testing if necessary and speed up treatment.