One of the most important characteristics of knitted fabrics is edge curling. Knitted fabric edges tend to curl to the fabric technical face or back. This tendency is due to release of energy that stored in loops by forces and bending and twisting moments which are applied to changing the straight yarn into a loop during knitted processing. Thus, when the fabric is cut, curling at the edges started with the discharging of this energy .In a dry-relaxed state, some forces act on the loop in a knitted fabric. These forces arise from tension acting on the structure during knitting, from jamming and also from changing the straight yarn into a loop by bending and twisting. Energy on the loop resulting from these forces causes a many problems such as shrinkage and curling. The yarn when bent into a loop during knitting tends by virtue of its elasticity, to straighten itself out and if the edge of the fabric is unrestrained, these forces in all the loops can be sufficient to lift the fabric and form it into a roll. Edge curling is important problems which occur, in particular, in plain knitted fabrics owing to the unbalanced yarn bending moment existing in the three-dimensional nature of the structure. The curling occurs at the upper and lower edges of a piece of fabric towards the front side and at the left and at the right edges of the fabric towards the back side. In a plain knitted fabric structure, the yarns want to adopt a straight form but it is prevented from doing so by neighboring loops. Thus, curling can start at the edges as there is no neighboring loop on one side to prevent curling. The edge curling can create some problems during the creation of plain knitted clothing goods. Edge curling of knitted fabrics, is one of their defects and results problems and waste in finishing and sewing, so study and measurement of curling is very important. In previous studies, researchers have tried to identify factors affecting curling and measurement curling distance by calculating moments on the loop that force fabric to curl. These moments are functions of fabric and yarn parameters such as yarn diameter and bending rigidity, loop length, wale-course spacing and fabric structure which are known before fabric production. It will be possible to calculate material loss duo to curling during garment preparation with predicting of it before production. The aim of this study is investigation into yarn and fabric parameters on curling of single jersey weft knitted fabrics. Single jersey weft knitted fabrics were produced on a circular knitting machine. Taguchi’s experimental designs were used to examine the effect of different fiber, yarn and fabric parameters such as fiber material and blend percent, yarn twist and count, fabric structure and density, relaxation time after cutting fabric and cut direction on fabric curling. The results show that the investigated parameters have significant effect on fabric curling in both wale and course cut direction. The factors were ranked due to their effect on the edge curling. According to the results of Taguchi method, fabric structure shows the strongest effect on fabric curling and fabric density is second in both wale and course cut direction. In wale-cut direction, relaxation time after cutting is third, yarn twist is fourth, blend percent is fifth and is followed by yarn count. In course-cut direction, yarn twist is third, yarn count is fourth, blend percent is fifth and is followed by relaxation time after cutting. The optimum conditions were determined through the experimental design. In both cut directions, the double cross tuck structure, the lowest density, the lowest yarn twist and 50/50 cotton- polyester yarn conclude minimum of edge curling. Measuring of edge curling immediately after cutting results lower curling than measuring it after relaxation time. In wale-cut direction, the medium level of yarn count (25Ne) and in the course-cut direction the lowest level of yarn count (20Ne) were identified as optimized levels. At the end of this study, the loop of plain structure was simulated in ABAQUS software and the model results as a curling index, compared with experimental results. Comparing the amount of stress in loop model with curling distance, due to ignoring friction forces in model, is impossible. For more accurate analysis, these stresses must compare with the edge curling forces.