Leaf area index (LAI), defined as half the total developed area of green leaves per unit ground horizontal area, is an extremely important vegetation characteristic parameter that describes the construction of vegetation canopy. For the past few years, LAI has been estimated operationally at a regional or even global scale by means of various retrieval methods using remotely sensed optical imagery. Understory vegetation (e.g., grasses, herbs, shrubs, etc.) is the layer of foliage below the forest canopy. As a background signal source in remote sensing, understory has had a serious impact on estimating many forest canopy parameters by remote sensing techniques for two reasons. on the one hand, it has very similar characteristics with those of the forest canopy because they are all vegetation. on the other hand, it has greater spatial-temporal heterogeneity than other backgrounds in remote sensing, e.g., soil, water, rock and litter, etc. The purpose of this paper is to determine the impact of understory on LAI inversion of forests with different canopy closures in different seasons. In this study, data from five field investigations and corresponding Chinese HJ-1 CCD remote sensing images were collected and analyzed to study the impact of understory on LAI inversion of Pinus massoniana during the period September 2012 to October 2013 in Chuzhou, Anhui province. By building empirical models of LAI of Pinus massoniana with different canopy closures and Normalized Difference Vegetation Indices (NDVI) in different seasons based on the phenology of the understory and comparing the difference of LAI of tree layers and the LAI of all layers of the ecosystem using a new way of LAI measurement, the impact of understory on the calculation of LAI of Pinus massoniana with different canopy closures and in different seasons was found. The results show that: (1) Understory had minimal impact on LAI inversion of Pinus massoniana in winter; R² of the linear relationship between NDVI and LAI was 0.65. Understory had the most serious impact in summer; R² of the linear relationship was only 0.25. The impact of understory in spring and autumn was greater than in winter and lower than in summer, R² of linear relationship was about 0.47. However, R² of linear relationship in the quadrats A, which were less affected by understory, was almost higher than 0.60 in all seasons (slightly less than 0.60 in summer). The reason was that the phenology of the understory caused different impacts in different seasons; (2) A significant difference between tree layer LAI and LAI of all layers in the same season was found, and the biggest gap was 2.93 and the maximum multiple was 2.45; (3) the R² of the linear relationship between LAI of all layers and NDVI was about 0.66, and R² of the logarithmic relationship was more than 0.68. However, the correlation between tree layers and NDVI was poor (R² was only 0.30).These findings indicate that the impact of understory on LAI inversion of Pinus massoniana can be calculated when relationships between the LAI of all layers and NDVI in winter and other seasons are determined. Finally, the difficulties of studying understory impact on forest LAI retrieval are discussed, and several suggestions are proposed for future studies.