This paper presents a feed forward neural network model for predicting the trajectory performance of an artillery rocket. Three problems of artillery rocket performance are addressed under known ambient conditions to predict 1) the range for the specified firing angle, 2) the firing angle for the specified range, and 3) the range obtainable under standard (specified) conditions. The neural model bypass the need for any mathematical modeling or their solutions as hitherto resorted to for predicting the trajectory performance. For the purpose of present study, data from the range table of G-rocket, supplied by the user agency, are used. An appropriate set of input/output variables from the given data is selected to train, validate, and predict via the proposed neural models. A comparison of the neural model estimate with the values from the range table shows close agreement for all the proposed neural models. I. Introduction rtillery forms an important wing of an army in providing firepower, during both war and cross-border skirmishes with the enemy. Artillery rockets are a class of projectiles around which much of the aeroballistic theory was originally developed, and it continues to form a significant part of the aero ballistician’s interest. The effectiveness of artillery rocket is largely judged by the accuracy in hitting the targets. Various error sources inherent in the rocket system, together with the external conditions such as wind, cause dispersion of the rocket from its intended path. The actual path traversed by the rocket is compared to the predicted trajectory in order to calculate its accuracy. Beginning with the simplest, but relatively inaccurate, in-vacuo trajectory mathematical model, more and more sophisticated models of increasing accuracy, such as the point mass model, the modified point mass model, and the six-degree of freedom model, 1 have been developed. However, even the best of these models have their limitations because of 1) an inability to model all of the problem variables (e.g., the initial conditions at the time of rocket leaving the launcher, the tip-off effect, the aerodynamic jump, the variable atmospheric conditions, etc.) and 2) the non-availability of accurate and reliable aerodynamic coefficients (e.g. drag coefficient, damping in roll derivatives, etc.) required as inputs by the mathematical models. The recent interest in evolving applications of artificial neural networks (ANN) to diverse fields such as signal processing, pattern recognition, robotics, medical diagnosis, system identification, and control have led many researchers to explore their capability for aerospace engineering problems. The neural modeling has been employed in solving aerospace problems such as aerodynamic modeling, 2 buffet, 3