This research was based on laboratory experiments for investigating the hydraulic characteristics of a rectangular weir with three rectangular bottom openings. Two cases were investigated ,case A of equal size of openings, and case B of different size of the mid opening than an equal size of the two side openings(. Fifty-six physical models were made for this structure with different geometrical dimensions. Experimental data were analyzed in order to find the discharge coefficient for three flow cases. Case 1 when the water level is under the openings height (partial flow through the openings only), case 2 when the water level is between the openings height and under the weir crest level (full flow through the openings only), and case 3 when the water level is higher than the weir crest level (flow through both openings and over the weir). The analysis of the experimental results was done by using an equation prepared for each flow case, to find the discharge coefficients. Dimensional analysis was made to relate the discharge coefficient with different geometrical and flow non dimensional variables for each flow case, using Artificial neural networks. The highest value of the discharge coefficient, was obtained for flow case number 3, (0.6188, while its lowest value was occurred for flow case number 1,( 0.0288).The correlation coefficients found between each of the flow cases discharge coefficient, and the ratio of individual opening width to the total head (bo/H) are the highest among the other relevant variables and have negative values (-0.942,-0.937,-0.871) for the three flow cases respectively for equal size and (-0.953,-0.941,-0.503. For all the cases tested the coefficients of variation for the discharge coefficient were in the range of (0.211 - 0.563) which indicate the necessity of obtaining a model to describe this variation. Six ANN models were developed herein using the (SPSS software, version 19) to express the discharge coefficient as a function of the different geometrical and flow non dimensional variables. Three of these models are for the three flow cases with equal size of openings, while the other three are for the three flow case for unequal size of openings. The models range of correlation coefficients is (r=0.999 - 0.997).