In order to study the influence discipline of parameters of gripping and delivering device (GDD) for corn harvester for reaping both corn stalk and spike in the cramping-cutting-transporting type on corn harvesting performance, the working principle of GDD was analyzed. Box-Benhnken design （BBD） response surface method was used to design the field experiments. An orthogonal rotary combination experiment was conducted with the angle of gripping chain, the teeth number of sprockets in enter shaft, the horizontal distances between reciprocating cutter and the nip point of gripping chains and the working speed of corn harvester as four independent factors, as well as loss rate of corncobs, maximum offset of x axis and y axis as three dependent indices, while using Zhengdan corn as the test materials. The height of 1.2m in corn straw was signed by colorful paper tape. And high speed cameras were utilized to record the motion of mark point. ProAnalyst motion analysis software was applied to get the maximum offset of x axis and y axis. Design-Expert software was employed to analyze the test results, mathematical equations between independent and dependent factors were gained. The results indicated that four independent variables had a second nonlinear relationship with loss rate of corncobs and maximum offset of x axis, and the horizontal distance between reciprocating cutter and the nip point of gripping chains had the greatest influence as well as the angle of gripping chain had the least effect, respectively. The four factors had no significant difference in maximum offset of y axis. The interaction of factors only had significant effect on loss rate of corncobs. At the angle of gripping chain of 19.96°, the teeth number of sprockets in enter shaft of 22.09, the horizontal distances of 22.33mm as well as the working speed of corn harvester of 1.31m/s, the GDD showed optimal performance with loss rate of 0.4% and maximum offset of x axis of 24mm. Field verification test was carried out after the combination of the optimal parameters, and the results showed that the regression model had good reliability to predict the loss rate of corncobs and maximum offset of x axis. According to comparison of transplanting performance before and after parameter optimization, loss rate of corncobs was reduced by 2.4 percentage points, which was superior to the technique indexes of national and industry standards as well. The research provided a basis on the design of gripping and delivering device of corn harvester for reaping both corn stalk and spike.