mart cities (SCs) aim to improve the quality of life, create a sustainable environment, promote economic development, enhance security, and meet user needs. Coupling SCs and Internet of Things (IoT) can enhance decision-making by combining different communication devices. A smart city uses IoT devices to gather data from various sensors such as GPS, sensing devices, and cameras. These sensors generate massive amounts of data. The Artificial intelligence (AI) tools can use this data to analyze and make better predictions. Machine learning (ML) is a subset of AI that uses data in smart cities to improve services to meet user needs in an automated manner. This paper proposes an ML framework for evaluating and improving the use of the smart cities' resources and services toward enhancing the life quality and increase economic development with high security. ML models such as random forest (RF), extra trees (ET), histogram gradient boosting (HGB), RF with bagging, ET with bagging, and HGB with bagging are employed in this study. Data preprocessing methods are applied to enhance the data quality. ML models are applied on five scenarios, including a transportation system, smart agriculture system, smart grid system, security system, and Industry 4.0 with IoT system, to evaluate and improve the services in smart cities. The ML models achieved higher accuracy and performance across these five scenarios. HGB obtained the highest accuracy of 99.99% in the transportation system; RF achieved the highest accuracy in smart agriculture systems, Industry 4.0, and security systems, with an accuracy of 99.79%, 99.54%, and 95.93%, respectively, while ET achieved the highest accuracy in the smart grid system, which was 96.87%. Based on this, we concluded that the proposed framework can improve the services of the smart cities to meet the growing needs of users.