Earthquake is a type of disaster that occurs suddenly in different magnitudes. When magnitude of an earthquake increases it is expected that the effects are much more. Earthquakes in varying magnitude between 4 Mw and 5 Mw cause uneasiness among the public even if they do not cause heavy damage. The aim of this study is to estimate the probability of an earthquake between 4.0 and 5.0 by using artificial neural network model. Monthly real data between 2006 and 2015 is used for the model. Data is analyzed in MATLAB neural network tool, then estimated output value obtained via analysis and output of test value is compared with regression equation. Besides, seasonal effects on magnitude of earthquake are examined. Results show that 90.51% of the earthquake probability between 4.0 and 5.0 can be estimated by using artificial neural network model.