This paper introduces a new class of stochastic neural networks with time delays in the leakage terms. A key characteristic of this new stochastic neural network is that its coefficients are dependent on expectations. We first establish a novel stability lemma for this new model. Then, by applying this new stability lemma, Itô׳s formula, Lyapunov–Krasovskii functional, stochastic analysis theory and matrix inequalities technique, we show that the suggested system is exponentially stable in the mean square. Moreover, some remarks and discussions are given to illustrate that the obtained results are significant, which generalizes and comprises those obtained in the previous literature. Finally, an example is given to show the effectiveness of the theoretical result.