In this paper, we propose a new simulation approach based on a cellular automata to predict the closing or opening of irrigation culvert. To solve this problem, water quality parameters such as salinity, temperature, pH, dissolved oxygen, etc. were measured at culverts. Then, opening or closing the culverts depending on the quality of the water there is considered. However, due to the large number of culverts, it is very time consuming to carry out manual measurements of all culverts. It is important to have a measure to help predict the water quality at culverts so as to reduce the amount of effort and time spent by farmers, meanwhile it helps farmers to feel secure to do the production (The simulations are based on data on water quality collected at culverts in subregion X - South Ca Mau, Ca Mau province, Vietnam).