This paper describes a method of estimating the limiting factors for the distribution of eelgrass beds using a neural network from the results of an environmental investigation. After conducting the environmental investigation in eleven areas of the Seto Inland Sea, we used the back propagation method to model a layered neural network to determine the degree of eelgrass growth under differing environmental conditions. The explanatory variables of this network are water depth, net photosynthetic rate, wave height, gradation of bottom material (the proportion of gravel, sand, silt and clay), and oxidation-reduction potential of bottom material. We performed a neural network simulation to estimate the limiting factors for the distribution of eelgrass beds along the coast near Kasaoka-shi, Okayama Prefecture. Our simulation assumed the following four cases about the input values for the environmental conditions at each prediction point in the simulation range from the results of the environmental investigation on this coast: 1) present environmental condition, 2) extinction coefficient (index of sea water turbidity) half of present condition, 3) wave height half of present condition, and 4) proportion of silt and clay half of present condition. For each case, the range of possible prediction points with eelgrass growth present was extracted as the potential area of eelgrass bed distribution. The limiting factors were estimated by comparing these results. The potential area of distribution was dramatically larger for case 2 than case 1. Case 3 increased the area slightly, with little or no increase in case 4. This simulation revealed that sea water turbidity, or loss of light, is the most powerful limiting factor, but wave height also limits the distribution of eelgrass beds on this coast. Given the results of an environmental investigation, this method is a simple, useful way to estimate the factors limiting the distribution of eelgrass.