Leaf area index (LAI) of three types of forests in the Qianyanzhou of subtropical China was investigated with fisheye photographs and CI-110 canopy analyzer. DBH-leaf area (hemi-surface) regressive model was used for assessing the measurement accuracy by photogrammetry and the canopy analyzer. Field survey was conducted on 74 plots. The values of LAI of for the Pinus elliottii forest, the Pinus massoniana forest, and the Cunninghamia lanceolata forest were 1.339 (0.71-1.79), 1.491 (1.11-1.92) and 1.984 (1.26-2.59) by fisheye photos photogrammetry, 1.355 (0.38-2.85), 1.265 (0.62-2.55) and 2.140 (0.76-3.31) by the CI-110 canopy analyzer, and 4.53 (2.24-6.34), 3.88 (1.19-6.28) and 6.63 (3.5-8.37) by DBH-leaf area model, respectively. Compared with the results by individual-based leaf area model, the two indirect measurement techniques underestimated the LAI, but there are correlations among the three methods. The individual-based model is considered the most reliable approach for estimating leaf area index but labor consuming. Hemispherical photography proved to be more accurate than the CI-110 canopy analyzer for estimating LAI in the planted forests of the subtropical China.