MMAUD is a comprehensive multi-modal anti-UAV dataset designed to address the challenges posed by small unmanned aerial vehicles (UAVs). It integrates stereo vision, LIDAR, radar, and audio arrays to provide a rich and diverse data source for advanced UAV detection, classification, and trajectory estimation. The dataset includes Leica-generated ground truth data with millimeter-level accuracy, enhancing the credibility of algorithms and models. It is the first dataset to offer a unique overhead aerial detection perspective, surpassing previous datasets in fidelity and applicability. MMAUD is cost-effective and adaptable, enabling users to experiment with new UAV threat detection tools. The dataset simulates real-world scenarios by incorporating ambient heavy machinery sounds, enhancing its applicability in proximate vehicular operations. It is expected to play a pivotal role in advancing UAV threat detection, classification, and trajectory estimation capabilities. The dataset, code, and designs are available at https://github.com/ntuaris/MMAUD. The dataset includes data from various drone types and ambient noise sequences, with each sequence containing visual, audio, LIDAR, and radar information. The dataset is accessible in two formats: rosbag and filesystem. It is designed for 2D detection and 3D estimation methods, with results showing the performance of various models. The dataset also includes ground truth data at a rate of 5Hz, which may not be sufficient for all research scenarios. The dataset faces challenges such as limited geographic coverage, sensor synchronization, limited drone variability, and missing data. Despite these challenges, MMAUD contributes valuable insights and datasets to the field of drone detection, tracking, classification, and trajectory estimation.MMAUD is a comprehensive multi-modal anti-UAV dataset designed to address the challenges posed by small unmanned aerial vehicles (UAVs). It integrates stereo vision, LIDAR, radar, and audio arrays to provide a rich and diverse data source for advanced UAV detection, classification, and trajectory estimation. The dataset includes Leica-generated ground truth data with millimeter-level accuracy, enhancing the credibility of algorithms and models. It is the first dataset to offer a unique overhead aerial detection perspective, surpassing previous datasets in fidelity and applicability. MMAUD is cost-effective and adaptable, enabling users to experiment with new UAV threat detection tools. The dataset simulates real-world scenarios by incorporating ambient heavy machinery sounds, enhancing its applicability in proximate vehicular operations. It is expected to play a pivotal role in advancing UAV threat detection, classification, and trajectory estimation capabilities. The dataset, code, and designs are available at https://github.com/ntuaris/MMAUD. The dataset includes data from various drone types and ambient noise sequences, with each sequence containing visual, audio, LIDAR, and radar information. The dataset is accessible in two formats: rosbag and filesystem. It is designed for 2D detection and 3D estimation methods, with results showing the performance of various models. The dataset also includes ground truth data at a rate of 5Hz, which may not be sufficient for all research scenarios. The dataset faces challenges such as limited geographic coverage, sensor synchronization, limited drone variability, and missing data. Despite these challenges, MMAUD contributes valuable insights and datasets to the field of drone detection, tracking, classification, and trajectory estimation.