We examined photographic methods and conditions for estimating the leaf area index (LAI) of the ‘Delaware’ grape using a digital camera with a fisheye lens. We separeted leaves from other parts in the pictures by binarizing the digital images and calculating the estimated LAI based on optical assumptions from the number of pixels. It was found that the optimum conditions for shooting hemispherical images were cloudy days. Furthermore, as the gap fraction changes significantly on clear days, it was found that the optimum conditions on these days were sunrise or sunset. Due to the nature of ‘Delaware’ grapes, which are trained to trellis heights of approximately 150 cm, it was determined that the estimated LAI more closely approached the measured LAI if images were taken from a distance of 100 cm to 150 cm from the trellis; however, the estimated LAI showed a trend toward overestimation. According to an examination into the cause of overestimation, the effect of the existence or non-existence of fruit was less than the effect of the shooting time. Because we could not discount the effect of the construction materials of the trellises, greenhouse, branches, and others, we added the effect to the gap fraction and recalculated the corrected estimated LAI. Using this correction, it was found that it is possible to estimate LAI faster and simpler than existing methods.