In ecological studies, it is a common situation occured that population density of species extremly increases or decreases in certain periods depending on many abiotic and biotic factors. Because of ecologial factors that cause high level fluctuation in population density, It is possible to get zero individual at samplings, and on the other hand, differences between maximum and minimum values obtained in different samplings intervals can be very high. Because this type of data based on counting does not show normal distribution, and shape of the distribution is skewed to the right because of the abundance of zero, using the Zero-inflated Poisson regression method (ZIGP) is required. This study was carried out to obtain information on effects of physico-chemical environmental conditions on population fluctuation of Notonecta viridis. Samplings were conducted with monthly periods along the coastal band of Van Lake in 2005-2006. Samples were taken from 20 sampling places where have three different characters as stream entrances, settlements and natural coastlines. Reults were analysed by using ZIGP regression model. According to results, Effect of sampling intervals and sampling stations on population densities of Notonecta viridis were important. On the other hand, HCO3 had negative effect on population densities in zero-inflated model while it had possitive effect on population densities in other two models. It was determined that Fe effected the species populations in the negative way in the mean regression model, and Cl and Mg effected it in possitive way in the overdispersion regression. In the result, it was deductived that Notonecta viridis was found excessive numbers or none in some sampling stations because of the pysico-chemical structures of water.