In this paper, the combination of finite element modeling and experiment was used to study the effects of different vibration parameters such as vibration time (A), amplitude (B) of the vibration excitation point on apricot tree vibration and vibration frequency (C), and optimize the performance of forest fruit vibration harvesting prototype. The effects of different vibration parameters on the vibration detection points of apricot trees were analyzed by vibration response test. The response analysis of apricot tree free modal vibration using ANSYS Workbench software showed that apricot trees had typical free modal responses at 15, 26, 30 and 40 orders, the corresponding frequencies were 5.5, 10.5, 15.1, and 29.3 Hz, respectively; and the maximum displacement deformation of fruit trees was 286.5, 261.9, 267.2, and 273.5 mm, respectively. The typical free modal cloud image analysis showed that the typical free modalities of the 15th and 26th orders were higher in the terminal end responsiveness of the apricot branches. The 30th order fruit tree had a high overall response consistency, and the apricot tree had the largest deformation response at the end; at the 40th order of 29.3 Hz, some branches of the fruit trees were severely deformed, and the overall structure of the tree was easily destroyed. The optimal response frequency of apricot tree vibration harvesting ranged from 0 to 20 Hz. The harmonic response analysis showed that in the optimal frequency range, the acceleration of the same position increased with the increase of the excitation amplitude at the same frequency, but the overall variation of the vibration curve was consistent with the trend; When the amplitude of the vibration excitation point of the apricot tree was 5, 10 and 15 mm, at the same frequency, as the excitation amplitude increased, the acceleration at the same position increased, but the overall variation of the vibration curve was consistent with the trend. The three-factor and three-level vibration response tests were conducted to study the effects of vibration time, vibration frequency and excitation point amplitude on the acceleration of three different detection points. The multivariate regression analysis of variance showed that the accelerations P1 and P2 of detection points 1 and 2 were less than 0.000 1, the overall model was highly significant (P<0.01), and the model regression terms B and C were significant. The acceleration P3 of the detection point 3 was 0.000 1, and the overall model was highly significant (P<0.01), except that the acceleration P3 value of the detection point 3 of the model regression terms AB, AC and A2 was not significant (P>0.05), and other regression terms were significant (P<0.05). The factors which affected the acceleration of the detection points 1 and 2 were the same as the amplitude of the excitation point, vibration frequency and vibration time. The order of the magnitude of the acceleration affecting the detection point 3 was the vibration time, the amplitude of the excitation point, and the vibration frequency. Through the response equations of the three detection points, the coefficient of determination R2 of the regression equations from the bottom to the top three detection points were 0.906 7, 0.879 3 and 0.973 3, respectively. Using the Design-Expert 10.0.3 software to optimize the detection point response equation, the optimal vibration recovery parameter combination was that the vibration time was 7.207 s, the vibration frequency was 15 Hz, and the amplitude of the excitation point was 10 mm. The verification test showed that the acceleration from the bottom to the top of each detection point was 10.4, 10.2 and 9.3 g, which was similar to the optimized value. These conclusions can provide design and theoretical basis for the design of mechanical parameters of apricot vibration harvesting.