•Introduce a trustworthy smart city service delivery solution at the edge of the Internet.•Improve the performance and the security aspects of smart city services.•Provide a detailed evaluation of the framework with support of machine learning.•Provide a detailed state of the art of smart city services supported by IoT and Edge computing. The proliferation of smart city and Internet of Things (IoT) applications has introduced numerous challenges related to network performance, reliability, and security. Moreover, the distributed nature of the smart city and IoT infrastructure hs led to issues in regards to service availability, reliability, sustainability and security. Edge computing provides a decentralized computing and communication framework for different types of applications such as intelligent transportation systems, cognitive assistance, health and social services. Edge computing helps in improving the performance of such applications and reduces the end-to-end latency incurred for such time-critical applications. In this article, we introduce a trustworthy smart city service delivery solution at the edge of the network. The solution uses a collaborative technique between distributed edge servers and privacy mediator nodes with the support of an intrusion detection system to enhance the availability, reliability and security of smart city applications. Simulation results show a reduction in the delay per service request by 39.2% for highly dense environments, and up to 62.6% for lightly dense environments. Moreover, the solution reduces the dropped service requests below 2% with high accuracy and detection rates and low false-negative rates.