Acrylic fibers are major part of synthetic fibers. Its special characteristics are unique, so that many researchers have been attracted to them. بهینه سازی و کنترل فرآیند تولید الیاف، تأثیر مستقیم بر روی هزینه، انرژی و زمان تولید دارد. تولید بیشتر با هزینه کمتر و با کیفیت بالا مسئله ای است کارخانجات تولید الیاف با آن روبه رو هستند. Process optimization and control of fibers have direct impact on the cost of energy and time. Produce more with less cost and high quality is a problem that manufacturing factories face it. طبیعت فرآیند معمولا بسیار پیچیده است و شامل پارامترهای زیادی است که هر کدام از آنها نیز تأثیر مستقیم بر روی عملکرد سیستم دارد.Nature of the process is usually very complex and includes many parameters. در طی چند سال گذشته برخی از محققین از تابع های چند متغییره برای بهینه سازی فرآیندهایی از قبیل پلیمرزاسیون برای کنترل مستقیم بر روی خط تولید استفاده کرده اند.During the past few years, some researchers have used multi-objective functions to optimize processes such as polymerization, but it needs somebody with high experience and skillful because in some cases there is need to user decision toward applied variables in functions. Recently computer methods such as genetic algorithms (GA) and neural networks (ANN) for optimizing and forecasting the behavior of chemical processes are used and the results have been remarkable. There is no comprehensive research about the optimization of acrylic fiber production in dry-spinning method using computer algorithms. Therefore, this study has tried to use the above mentioned methods. To predict the behavior of dry-spinning process many parameters such as temperature of extruder in the head and around, solution viscosity, water percent, the amount of formic acid solution and remaining time of solution in the reactor have been measured. Regarding to color index of manufacturing fiber as an indicator of production quality and using statistical methods, the affecting parameters on the process have been determined. After that, an ANN was designed to predict the color index. Then the parameters of ANN have been optimized using GA and GA itself has been optimized by tryial and error method. Finally, an ANN with high accuracy to predict dry-spinning process was designed.