Son yyllarda Ya?y?-Aky? ili?kisinin modellenmesinde, giderek daha farkly alanlarda kullanylmaya ba?layan ve oldukça ba?aryly sonuçlar veren, ancak su mühendisli?i alanynda henüz yeni yeni ivmelenen Yapay Sinir A?lary, Genetik Algoritma ve Bulanyk Mantyk gibi Yapay Zeka Yöntemlerinin kullanymy gündeme gelmi?tir. Bu çaly?mada, Orta Fyrat Havzasynda bulunan 2122 numaraly akym gözlem istasyonuna ait aylyk ortalama akym verileri ile 17099 numaraly ya?y? gözlem istasyonuna ait aylyk toplam ya?y? verileri arasyndaki ili?ki yapay sinir a?lary metotlaryndan Yleri Beslemeli Geri Yayynym Sinir A?y (YBGYSA) metodu ile ara?tyrylmy? ve ardyndan bu sonuçlar daha klasik bir yöntem olan Çoklu Do?rusal Regresyon (ÇDR) yöntemi ile kar?yla?tyrylmy?tyr. Akym ve ya?y? verilerinin girdi olarak kullanyldy?y bu i?lemde yeni akym de?erleri tahmin edilmi?tir. Çaly?ma sonunda YBGYSA yönteminin ÇDR yöntemine göre daha iyi sonuç verdi?i görülmü?tür

Artificial intelligence methods such as Artificial Neural Network (ANN), Genetic Algorithm (GA) and Fuzzy Logic (FL) provide prosperous results have recently been used in the modelling of rainfall-flow relations and they are becoming more popular in hydraulic engineering practices. In this paper the relations between the average monthly flow data from the flow observation station numbered as 2122 and the monthly total rainfall data from the rainfall observation station numbered as 17099 located in the central Euphrates river basin are investigated by using the feed-forward back-propagation neural network (FFBPNN) method from ANN solutions and afterwards the results are compared using Multi linear Regression (MLR) method. New flow values are estimated by this procedure that uses the flow and rainfall data as input. This paper concludes that FFBPNN method provides better results compared to the results from MLR method.