The exponential synchronization for a class of neural networks (NNs) based on dynamic event-triggered mechanism (DETM) is researched in this article. Firstly, unbounded distributed delay is introduced into the NNs. Next, based on the characteristics of sawtooth structure, an improved bilateral Lyapunov-Krasovskii functional (LKF) is constructed, which involves more information. By using improved integral inequality, some sufficient conditions are achieved for the exponential stability of the synchronization error system. Due to the influence of external factors or internal components, the controller parameter is uncertain. Then, a non-fragile controller is designed based on the decoupling technique. Moreover, a co-design scheme of controller gain and event-triggered matrix is obtained based on linear matrix inequality technique. Finally, two examples are used to illustrate the validity and feasibility of the presented method.