Code smells are indicators of poor software design or implementation choices that hinder software maintainability, performance, and overall quality of a software. Manual identification of code smells is time consuming and resource intensive. Detection tools have their own drawbacks including false positives, scalability, lack of context, lack of coverage, and poor usability. Furthermore, existing code smell visualization models are limited with abstraction levels, scalability issues, reliance on manual processes, and lack of empirical studies, which limit developers’ ability to make informed decisions efficiently. To fill this gap, this paper introduces a novel 3D visualization model that combines static analysis with `island` and `city` metaphors to represent classes, methods, and their relationships, along with associated code smells. The model was applied to a software project, and the usability of the model was evaluated in a pilot study. The findings from this study demonstrate that the proposed model provides an intuitive visualization of code smells than traditional tools, thereby supporting developers’ decision making and improving program comprehension.