In response to the problem of side buds being unable to be picked during the mechanized picking process of famous tea, an end effector was designed based on the relevant parameters of top buds, side buds, and tea stems, combined with the tea garden environment, which used the bending deformation of the cutter teeth of the split cutter to adapt to the interference of tea stems to pick side buds. The influencing factors of picking success rate were obtained by finite element simulation of cutting lateral buds: cutter tooth width, cutter tooth length and cutter thickness; the central composite design with three factors and three levels and response surface analysis were used to study the interaction of various factors on the success rate of picking; taking the picking success rate as the response value, a quadratic regression model was established to determine the significant primary and secondary order of the influence of each factor on the picking success rate as follows: cutter tooth length, cutter tooth width, and cutter thickness. Taking the picking success rate as the goal, the experimental factors were optimized, and the optimized parameters were obtained: the cutter tooth width, cutter thickness, and cutter tooth length were 2.6mm, 0.90mm, and 20.0mm, respectively. The optimized parameters were tested in tea garden picking, the results showed that the end effector could effectively pick tea leaves, and the success rate of picking top buds and side buds was 93% and 63%, the relative error between the experimental value and the predicted value was less than 5%, the optimized model was reliable.