In order to provide optimal theoretical cutting parameters of ramie stalk for ramie harvester, the cutting test of ramie stalk was conducted on the self-designed test bench in this paper.The test studied the influences of different geometrical parameters of blade, cutting speed, and ramie stalk feeding speed of reciprocating single movable blade cutter and double movable blades cutter on the cutting performance.The 4 groups of common blades had 2 blade lengths (81 mm, rice & wheat harvester blade; 120 mm, corn harvester blade) and 2 blade edge types(smooth-edge blade and serrated-edge blade), and the blade widths of various groups were the same(76 mm, equal to the driving distance of cutter bar on the test bench).Therefore, 5 factors were designed in the test: cutting speed, stalk feeding speed, blade edge type, blade length, and number of movable blades.Evaluation indices determined in the test were cutting power and number of failed stubbles (not cut off or broken); the smaller the 2 indices were, the better the performance would be.Besides, a multi-index evaluation model was established in the test and the values of multiple indices were reflected by the comprehensive score.The higher the score was, the better the performance would be.Among the 5 factors, blade edge type, blade length, and number of movable blades were obtained through model selection, and they were fixed at 2 levels.Cutting speed and stalk feeding speed were continuous values and the optimal values could be selected within a scope.Therefore, the optimal parameters of the 3 factors at 2 levels were determined by adopting the method of multi-factor orthogonal tests, and then the optimal parameter combination was obtained through the quadratic regression orthogonal rotation tests taking the other 2 factors i.e.setting cutting speed and stalk feeding speed as experimental factors.In the multi-factor orthogonal tests, the factors with significant influences on cutting power were cutting speed(P<0.01), number of movable blades(P<0.01) and blade length (P<0.1) according to the significant degree.The combination of the optimal level was double movable blades cutter and short blade(81 mm) under slow cutting speed(1 m/s).In addition, stalk feeding speed and blade edge type had no significant influence on cutting power.Meanwhile, the factors with significant influences(P<0.01) on the number of failed stubbles were blade edge type, blade length, cutting speed, and stalk feeding speed according to the significant degree.The combination of the optimal level was serrated-edge blade and long blade(120 mm) under fast cutting speed(1.2 m/s) and slow stalk feeding speed(0.7 m/s).In addition, the number of movable blades had no significant influence on the number of failed stubbles.Besides, the factors with significant influences on comprehensive score were blade edge type(P<0.01), blade length, number of movable blades, cutting speed and stalk feeding speed(P<0.05 for the latter 4 factors) according to the significant degree.The combination of the optimal level was serrated-edge blade, long blade(120 mm) and reciprocating double movable blades cutter under slow cutting speed(1.0 m/s) and slow stalk feeding speed(0.8 m/s).Furthermore, according to the combination of optimal score in the orthogonal tests, this paper conducted the quadratic regression orthogonal rotation tests for the 2 continuously adjustable factors i.e.cutting speed and stalk feeding speed, investigated their influences on cutting power, cutting quality and comprehensive score as well as their interaction effect, and established the quadratic regression functions respectively.In terms of the cutting power, when the cutting speed was 0.878 9 m/s and the stalk feeding speed was 0.862 4 m/s, the 1 m-swath cutting power was 281.408 4 W, which was the minimum value.As for the cutting quality, when the cutting speed was 1.161 4 m/s and the stalk feeding speed was 0.711 7 m/s, the number of failed stubbles in 1 m2 area was 5.691 1, which was the minimum value.For the comprehensive score, when the cutting speed was 1.092 0 m/s and the stalk feeding speed was 0.722 9 m/s, the score was 86.718 0, which was the highest value.Finally, according to the test results, the optimal parameter combination of ramie cutting test in theory was: cutting speed 1.092 0 m/s, stalk feeding speed 0.722 9 m/s, reciprocating double movable blades, serrated-edge blade, and blade length 120 mm, and under the condition, the 1 m-swath cutting power was 318.814 5 W, the number of failed stubbles in 1 m2 area was 6.006 4.In conclusion, the optimal parameter combination of ramie cutting adopted in the experiment is reasonable, and the experimental analysis results can be applied into subsequent related researches.