Maize (Zea mays L.) is a globally significant crop that plays a crucial role in feeding the growing global population. Among its various traits, plant height is particularly important as it affects yield, lodging resistance, ecological adaptability, and other important factors. Traditional methods for measuring plant height often lack cost-efficiency and accuracy. In this study, we employed a light detection and ranging (LiDAR) sensor mounted on an unmanned aerial vehicle (UAV) to collect point cloud data from 270 doubled haploid (DH) lines. This innovative application of UAV-based LiDAR technology was explored for high-throughput phenotyping in maize breeding. We constructed high-density genetic maps and assessed plant height at both single-plant and row scales across multiple developmental stages and genetic backgrounds. Our findings revealed that for many varieties and small areas, single-plant-scale estimation accuracy was superior to row-scale estimation, with an R² of 0.67 versus 0.56 and an RMSE of 0.12 m vs. 0.17 m, respectively. Two high-density genetic maps were constructed based on SNP markers. In Sanya and Xinxiang, the F1DH and F2DH populations identified 12 and 20 QTLs (quantitative trait loci) for plant height, respectively. The study successfully identified and validated QTLs associated with plant height, revealing novel genetic loci and candidate genes. This research highlights the potential of UAV-based remote sensing to advance precision agriculture by enabling efficient, large-scale phenotyping and gene discovery in maize breeding programs.