<p>This paper examines the impact of leverage heterogeneity on the stability of bank-asset bilateral financial networks. Based on the network framework proposed by Caccioli et al<i>.</i> (Journal of Banking &amp; Finance 46:233-245, 2014), we relax the assumption of homogeneous leverage among all banks by treating leverage as a random variable with its standard deviation serving as a measure of leverage heterogeneity. In simulations, randomly generated leverage values are assigned to banks according to different rules, resulting in various types of networks. The results reveal that if the mean of leverages is fixed at an intermediate level, leverage heterogeneity is inversely correlated to systemic risk when the average degree is small, but positively correlated when the average degree is large. If the mean of leverages is quite small or large, systemic risk is almost unaffected by the level of leverage heterogeneity and remains at a relatively high or a quite low level respectively. Additionally, the approach to stabilize the network through adjusting leverage heterogeneity also varies depend on the type of network and initial external risk. These findings are further confirmed by subsequent regression analysis which provides in-depth insights into our conclusions. Overall, this paper has significant implications and serves as a valuable reference for financial network risk control.</p>