Providing a convincing prediction for the urban future remains a challenge in planning research, discipline development and spatial practices. This paper reviews the progress in the research practice of urban models, evaluates the mainstream methods (i.e., macro, micro and integrated models) and highlights their strengths and weaknesses. Based on the identified trends in socioeconomic development and technological innovation, the challenges and opportunities of urban modelling are summarised against the backdrop of emerging big data, evolving technologies and behaviours, diversified policy needs, and growing complexity sciences. Furthermore, the future research directions of urban modelling are proposed, including the refinement of models, representation of technical and behavioural changes, model integrations across spatial scales and urban sectors, and ex-post evaluations and model updates. The findings from this research will help promote urban development theories and planning practices, since the improvements in urban models can establish a more convincing evidence base to address urban issues and support plan-making and evaluation.