The most important part of every electro-hydraulic servo system is its servo valve. A servo valve consists of very precise and sensitive components. A small change in the dimensions, metallurgical characteristics and other parameters of these components can produce instability and error in the performance of the system. In the design phase, the values of these parameters is considered as a crisp and certain numbers. Nevertheless, due to manufacturing limitations, the final value of these parameters always contains some tolerance and uncertainty. These uncertainties in the system parameters, creates a variance in the performance of the system from the ideal circumstances. In practice, it is possible that among the tens of thousands of produced valves, some cannot satisfy the standard specifications and pass the quality control. Therefore, initial evaluation of the effect of fabrication methods of the valve’s components on the statistic performance of the system can help the manufacturer to improve the reliability of its productions and decrease its expenses. Reliability can be used with the other factors like cost and accessibility, as a useful criterion for comparing various methods and techniques of manufacturing. In this research, a novel analytical method has been introduced in order to analyze the reliability of electro-hydraulic systems based on the manufacturing originated uncertainties existing in it. In this method, the number of failures is predicted using Monte Carlo simulation technique. For this aim, the dynamic model of the system was developed in the first stage. Afterwards, the uncertain parameters of the system were determined and based on the techniques and degree of precision used for the production of each valve’s component, the achievable tolerances were estimated for each parameter. By using a random number generator, each time, a set of samples was selected from these numerical ranges and substituted for the parameters of the dynamic model of the system. Simulations were then performed for each set of random values for thousands of times until a meaningful number of failures were observed in the performance of the system. In the next stages, by changing the manufacturing techniques and improving the achievable tolerance of each parameter in each stage, the simulations were performed again and the various cases were compared with each other. Simulations showed that the most important parameters in the frequency response of the system are length of air gaps in the torque motor and the diameter of flapper nozzles. By changing the manufacturing process of the valve components which affect the value of these two parameters and improving their normal distributions, the number of system failures was decreased significantly. On the other hand, by changing the manufacturing processes of some parts like spool and sleeve and improving their dimensional precision, no significant improvement was observed in the dynamic performance of the system. The simulations also revealed that the length of spool lands and the width of sleeve slots play the main role in the internal leakage of the servo valve and by changing the machining technique and improving the dimensional tolerances of these two items, the number of systems failures drop significantly. It was also found that, the clearance between spool and sleeve is not the dominant factor in the leakage of the valve. Keywords: Reliability Analysis, Manufacturing Uncertainty, Monte Carlo Simulation Method, Electro-Hydraulic Servo System, Servovalve