<p>In this paper, Parkinson oscillation in an improved thalamus - basal ganglia neural (Th-BG) model with discrete and distributed time delays is studied. The central medial - parafascicular complex (CM/PF), a significant target for deep brain stimulation (DBS), is considered in the model. The stability of the equilibrium point under both strong and weak gamma kernels is investigated to derive the conditions for the occurrence of Hopf bifurcation. At the same time, the impact of the gamma kernel function on system stability, bifurcation behavior, and neuronal discharge patterns are explored. The critical time delays for firing onset in both strong and weak gamma kernels are obtained. It is found that the strong kernels are more conducive to the generation of oscillations than the weak kernels. Moreover, the distributed self-feedback delays in both kernel clusters enhance the oscillatory behavior, especially in the strong kernel. Regardless of the kernel strength, controlling the initial changes in the distributed self-feedback delay of the globus pallidus (GP) nucleus and limiting the increase in the distributed self-feedback delay of the CM/PF can effectively prevent oscillations. Decreased excitatory feedback from the subthalamic nucleus (STN) to the CM/PF can amplify the stimulatory effect of the CM/PF-STN pathway on the STN-GP loop, thereby increasing firing rates. Within specific thresholds, the CM/PF can enhance oscillation and increase amplitude; however, exceeding these thresholds results in an absolute stable state of the system. Therefore, adjusting distributed self-feedback delays and modulating the excitatory influence of the CM/PF on the STN could serve as mechanisms to suppress abnormal oscillations associated with Parkinson’s disease.</p>