This paper introduces MFTraj, a map-free, behavior-driven trajectory prediction model for autonomous driving. The model leverages historical trajectory data and a novel dynamic geometric graph-based behavior-aware module to capture complex interactions in dynamic traffic scenarios without relying on high-definition (HD) maps. At its core, an adaptive structure-aware interactive graph convolutional network captures both positional and behavioral features of road users, preserving spatial-temporal intricacies. Enhanced by a linear attention mechanism, the model achieves computational efficiency and reduced parameter overhead. Evaluations on the Argoverse, NGSIM, HighD, and MoCAD datasets demonstrate MFTraj's robustness and adaptability, outperforming numerous benchmarks even in data-challenged scenarios without the need for additional information such as HD maps or vectorized maps. It maintains competitive performance even in scenarios with substantial missing data, on par with most existing state-of-the-art models. The results and methodology suggest a significant advancement in autonomous driving trajectory prediction, paving the way for safer and more efficient autonomous systems. MFTraj integrates four key components: a behavior-aware module, a position-aware module, an interaction-aware module, and a residual decoder. These components work together to analyze and interpret various inputs, understand human-machine interactions, and account for the inherent uncertainty and variability in the prediction. The model is evaluated on multiple datasets and shows superior performance in trajectory prediction accuracy and efficiency without additional map information. It also demonstrates strong performance even with significant missing data, highlighting its resilience and efficiency in predicting future vehicle trajectories.This paper introduces MFTraj, a map-free, behavior-driven trajectory prediction model for autonomous driving. The model leverages historical trajectory data and a novel dynamic geometric graph-based behavior-aware module to capture complex interactions in dynamic traffic scenarios without relying on high-definition (HD) maps. At its core, an adaptive structure-aware interactive graph convolutional network captures both positional and behavioral features of road users, preserving spatial-temporal intricacies. Enhanced by a linear attention mechanism, the model achieves computational efficiency and reduced parameter overhead. Evaluations on the Argoverse, NGSIM, HighD, and MoCAD datasets demonstrate MFTraj's robustness and adaptability, outperforming numerous benchmarks even in data-challenged scenarios without the need for additional information such as HD maps or vectorized maps. It maintains competitive performance even in scenarios with substantial missing data, on par with most existing state-of-the-art models. The results and methodology suggest a significant advancement in autonomous driving trajectory prediction, paving the way for safer and more efficient autonomous systems. MFTraj integrates four key components: a behavior-aware module, a position-aware module, an interaction-aware module, and a residual decoder. These components work together to analyze and interpret various inputs, understand human-machine interactions, and account for the inherent uncertainty and variability in the prediction. The model is evaluated on multiple datasets and shows superior performance in trajectory prediction accuracy and efficiency without additional map information. It also demonstrates strong performance even with significant missing data, highlighting its resilience and efficiency in predicting future vehicle trajectories.