People’s daily travel choices collectively shape traffic patterns, land-use efficiency, and broader urban systems. Traditional travel-behaviour models (TBMs) have long supported congestion management, infrastructure planning, and policy analysis. However, their adaptability to rapidly evolving technologies such as autonomous vehicles, connected transport, and smart mobility remains insufficiently explored. This study presents a systematic literature review of TBM research published between 2016 and 2025, adopting a PRISMA-guided and reproducible methodology to synthesise 41 peer-reviewed studies. The review analyses modelling paradigms, data sources, analytical techniques, and application contexts to identify dominant trends and structural limitations in the field. The synthesis reveals persistent gaps, including limited capacity to process real-time, high-granularity data, insufficient treatment of behavioural heterogeneity, and challenges in integrating emerging autonomous mobility scenarios within policy-relevant frameworks. Cross-cutting concerns related to privacy, equity, interpretability, and transferability are also evident across modelling approaches. Based on a structured qualitative and frequency-based synthesis, the study outlines priority research directions, including hybrid behavioural–machine learning models, dynamic calibration using crowd-sensed and multi-source data, explicit modelling of behavioural diversity in autonomous mobility adoption, and the development of standardised, privacy-aware multimodal datasets to enhance robustness and policy relevance.