A neural network is developed for the determination of leaky confined aquifer parameters. Leakage into the aquifer takes place from the storage in the confining aquitard. The network is trained for the well function of leaky confined aquifers by the back propagation technique and adopting the Levenberg–Marquardt optimization algorithm. By applying the principal component analysis (PCA) on the adopted training input data and through a trial and error procedure the optimum structure of the network is fixed with the topology of [2×10×2]. The network generates the optimal match point coordinates for any individual real pumping test data set which are incorporated with Hantush’s analytical solution and the aquifer parameter values are determined. The performance of the network is evaluated by real field data and its accuracy is compared with that of the type curve matching technique. The network eliminates graphical error inherent in the type curve matching technique and is recommended as a simple and reliable alternative to the type-curve matching technique.