The frame of the cotton picker is exposed to complex and varying loads during its operation. Therefore, conducting research on the vibration characteristics of the frame is crucial. In this study, vibration tests were conducted on the main vibration sources in a cotton picker at several measuring points on the frame. An accelerometer sensor was utilized to collect the signals. Fourier analysis was applied to analyze the vibration sources, encompassing the excitation frequency and the vibration source-coupled excitation frequency. Modal tests were also conducted to validate the finite element model and determine the natural frequencies of the frame. The results showed that the natural frequencies of the frame, specifically the third-order, fourth-order, and sixth-order frequencies, were comparable to the vibration source-coupled excitation frequencies. To prevent frame resonance, the response surface method was used to optimize the frame. Based on the MOGA algorithm, scheme 4 was identified as the optimal design. Furthermore, fatigue life calculations were carried out to optimize the parts with short lifespans on the frame, thereby enhancing the working performance.