In order to address the problem of tracking multiple low observable targets with multiple sensors, a superimposed intensity multisensor Gaussian mixture probability hypothesis density (SIM-GM-PHD) filter is proposed, together with a two-step extraction strategy. Firstly, a gate technique is used to choose valid sensors for each Gaussian component of the prediction intensity function. Then, each Gaussian component is updated with measurements from valid sensors, and the fusion posterior intensity is obtained by summing all local posterior intensities, as well as the global miss-detection intensity. Finally, a two-step strategy is proposed to estimate the number and states of the targets. Simulation results show the effectiveness of the proposed method.