Abstract: Leaves are the main organs of photosynthesis and transpiration in green plants. Leaf Area Index (LAI) is an important indicator reflecting light utilization and the structure of plant canopies. With advancements in remote sensing technology, the remote sensing inversion of LAI based on physical models has become an essential method for monitoring crop growth and estimating yield. Typically, it is assumed that the crop canopy structure is uniformly distributed and the one-dimensional radiation transfer model (1D-RTM), such as PROSAIL, is used for inversion. However, two issues are often overlooked: Is the crop canopy strucutre truly uniform distribution, and if not, how does this assumption affect the inversion results? In this study, we take maize, an important grain crops, as an example. Initially, 3D scenes of maize canopies are reconstructed. We quantitatively compare the reflectance simulated by PROSAIL model (1D-RTM) and LESS model (3D-RTM) to reveal the deviation in simulated reflectance caused by the non-uniform distribution of leaves during different growth stages of the maize canopy. Considering the variation characteristics of clumping index throughout the growth process of maize canopy, we use LESS model to calculate the corresponding gap fractions and clumping indices for different typical maize canopy structures, and establish a relationship between the clumping index and the effective leaf area index (LAIe). Furthermore, we use the clumping index to correct the vegetation LAIe obtained from PROSAIL model, yielding the actual LAI. The results demonstrate that the accuracy of the corrected LAI has significantly improved, with R2 increasing from 0.27 to 0.55. This method is expected to enhance the precision of crop LAI inversion using high-resolution remote sensing data.