Complex network reconstruction is a hot topic in many fields. A popular data-driven reconstruction framework is based on lasso. However, it is found that, in the presence of noise, it may be inefficient for lasso to determine the network topology. This paper builds a new framework to cope with this problem. The key idea is to employ a series of linear regression problems to model the relationship between network nodes, and then to use an efficient variational Bayesian method to infer the unknown coefficients. Based on the obtained information, the network is finally reconstructed by determining whether two nodes connect with each other or not. The numerical experiments conducted with both synthetic and real data demonstrate that the new method outperforms lasso with regard to both reconstruction accuracy and running speed.