The rapid urbanization of cities has increased the demand for smarter, more efficient infrastructure solutions. Theintegration of the Internet of Things (IoT) in urban environments has enabled real-time data collection fromsensors across various sectors such as traffic management, energy consumption, public safety, and environmental monitoring. However, the challenge lies in transforming this vast amount of data into actionable insights for bettercity management. This research addresses this problem by applying predictive analytics to IoT-driven smart cityapplications. We propose a framework that combines data preprocessing techniques with advanced machinelearning algorithms, including regression models and time-series forecasting, to predict key urban trends like trafficcongestion, energy demand, and air quality levels. Our methods have been tested on real-world IoT datasets from asmart city, achieving significant improvements in prediction accuracy compared to traditional approaches. Theresults demonstrate the potential of predictive analytics to not only improve operational efficiency but also toanticipate challenges before they arise, leading to more sustainable and responsive urban environments. This workhighlights the transformative role predictive analytics can play in optimizing IoT data for enhanced decision-makingin smart cities, offering valuable insights for urban planners, city authorities, and policymakers.