The goal of this research is to explore and analyze historical crime data in order to develop predictive models that can forecast future crime trends. By leveraging time-series analysis techniques, this study aims to uncover temporal patterns and identify key trends in criminal activities. Specifically, we use a variant (SARIMAX) of the well-known ARIMA (AutoRegressive Integrated Moving Average) model for time-series forecasting and analyze the prediction accuracy. The crime dataset (for the city of Chicago) used for this project contains detailed records of criminal incidents, including variables such as time, location, and crime type. The SARIMAX model helps to identify trends, seasonality, and periodic fluctuations in crime rates, while other machine learning techniques could be employed to optimize predictions based on additional factors such as geographic hotspots and socio-temporal variables. The relative error in the prediction is observed to be 2.88% to predict the number of crimes in a month for the city of Chicago whose average monthly crime rate is 28,677.