Subthalamic nuclei deep brain stimulation (STN-DBS) is a well-established treatment for Parkinson's disease (PD). Some studies have confirmed the long-term efficacy is associated with brain connectivity; however, whether the initial outcome is associated with brain connectivity and efficacy of prediction based on these factors has not been well investigated. In the present study, a total of 98 patients were divided into a training set (n = 78) and a test set (n = 20). The stimulation and medication responses were calculated based on the motor performance. The functional and structural connectomes were established based on a public database and used to measure the association between stimulation response and brain connectivity. The prediction of initial outcome was achieved via a machine learning algorithm-support vector machine based on the model established with the training set. It was found that the initial outcome of STN-DBS was associated with functional/structural connectivities between the volume of tissue activated and multiple brain regions, including the supplementary motor area, precentral and frontal areas, cingulum, temporal cortex, and striatum. These factors could be used to predict the initial outcome, with an r value of 0.4978 (P = 0.0255). Our study demonstrates a correlation between a specific connectivity pattern and initial outcome of STN-DBS, which could be used to predict the initial outcome of DBS.