Leaf area index (LAI) provides significant information about crop growth conditions, health status, as well as crop yield, and is of great interest for a variety of agricultural applications. With improved spatial, spectral and temporal resolution, data collected using recently emerged Unmanned Aerial Vehicle (UAV) platforms, have promoted remote sensing applications, particularly in field-based high-throughput plant phenotyping and precision agriculture. In this study, we investigate the potential of UAV-based hyperspectral and LiDAR data for sorghum LAI estimation using machine learning methods in the context of high-throughput phenotyping. Hyperspectral imagery-extracted canopy spectral features (i.e., vegetation indices) and LiDAR point cloud-derived canopy structure metrics, and their combination were used to estimate sorghum LAI using Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), Support Vector Regression (SVR) and Deep Neural Network (DNN). The results show that: (1) Integrated with different sensors, UAV is a promising platform for LAI estimation and crop growth monitoring in the context of high-throughput phenotyping; (2) both canopy spectral information and structure features are important indictors for sorghum LAI estimation, while combing spectral information with structure features provides improved estimation results; and (3) DNN model yields a promising prediction accuracy compared to other machine learning methods. UAV-based multi-sensor data fusion employed in this research delivers important insight for high-throughput crop growth monitoring and phenotyping at field scale.