Drawframe is a very critical machine in the spinning process. Its influence on quality, especially on evenness is very big. If drawframe is not set properly, it will also result in drop in yarn strength and yarn elongation at break. So the quality of yarn depends on the draw frame capability. To improve its performances, an autoleveller has often been used on the advanced drawframe.The basic method is to keep the output sliver density stability by adjusting the draftingratio continually according to the thickness of sliver fed into the drafting zone.In mechanical part, output part of drive system of the draw frame include front Roller,compress Roller, coiler head and cans remain the same equipment withoutautoleveller. Draft part, checking Roller and gantry creel drive part was re-designed.In the design, speed of front Roller kept the same, speed of middle and back Rollerwas changed at the same time, thus realized autolevel. The method was convenient fordesigning, guaranteed stability of output sliver and overcame the disadvantage thathard to satisfy requirements of system inertia when changed the speed of front Roller.In the control system part, an ideal and more applicable control strategy is put forward.Because the modeling drafting process is hard to be obtained and the high speed isdemand in practice, the linear drafting equation in the idea condition, in which onlythe delivery sliver deviation is calculated, has been adopted widely in present.However, the experts’ experiences concluding the deviation, the deviation change andthe stability demand of drafting may seem to reflect the process of the autolevellermore truly which establish the base of fuzzy controller’s application in this field. Atthe same time, for the difficulty of obtaining the full and precise control rules, byoptimizing the structure and the parameters of a basic fuzzy controller in severalaspects via neural network and gene algorithm, a new controller with quickerresponse and better performance is obtained. Finally, a set of autoleveller digitalcontrol scheme is designed and realized on the base of the serial communicationamong PLC, IPC and servo controller. Through the analysis and the compare of 4controllers, the optimized controller is approved to be an effective method in theprocess of autoleveller, and thus provides a new way in this field.Chaos non-linear time series method was used to discuss chaos phenomenon in sliverirregular data. As an important branch of non-linear science, chaos had showed anexciting prospect in scientific research and engineering technology. Chaos non-lineartime series unified certainty and random of the system and was more accurate to describe irregularity of sliver compared to traditional method. Fractional dimension, biggest Lyapunov index, time delay and inserted dimension number were first calculated;Evaluation of fractional dimension was obtained by GP saturated correlation dimension number method and biggest Lyapunov index was calculated by Rosenstein algorithm. Lap-correlation and inter information contend method were used to calculate delay time of phase space, inserted dimension number was calculated by saturation correlation dimension number method. The results showed that there was plus Lyapunov index and fractional dimension in the series, which meant the series, had the feature of chaos.The optimization of drafting system is another important aspect to improve sliver evenness. The motion of floating fibers during drafting has been considerably studied with the aim to control the floating fibers effectively, and to improve the sliver regularity. But for the speed change point distribution, it has been explained only qualitatively, especially for the pressure bar drafting system. This paper tries for the first time to establish a mathematical model of floating fiber motion through analyzing the forces acting on it, and simulate the fiber movement under various conditions by computer simulating calculation.The results can show the speed change point distribution with different statistical characteristic under various processing conditions, which had been obtained before only through experiments. In addition, the effects of frictional force distribution, fiber length, roller distance and drafting ratio on the speed change point distribution are analyzed respectively from the simulation results.It provides a new way for studying the movement of floating fibers in the drafting zone and for designing newly drafting system and optimizing the press bar system.This thesis focused on the theme of improving sliver evenness engage deeply study from transmission scheme, control method and drafting type to settle the disadvantage of FA306 to achieve the high-quality requirements of the textile