This paper investigates the problem of low probability of intercept (LPI)‐based adaptive jamming waveform design for distributed multiple‐radar architectures. Such a smart jammer system adopts a multibeam working mode, where multiple simultaneous jamming beams are synthesized to interfere with multiple radars independently. The primary objective of the smart jammer is to minimize the total jamming power by optimizing the transmitted jamming waveform while the achieved signal‐to‐interference‐plus‐noise ratio (SINR) and mutual information (MI) between the received echoes from the target at each radar receiver and the target impulse responses are enforced to be below specified thresholds. First, the expressions of SINR and MI are derived to characterize target detection and characterization performance, respectively. Then, two different LPI‐based jamming waveform design strategies are proposed to minimize the total noise jamming power by optimizing the jamming waveform while the achieved SINR/MI is enforced to be below a certain threshold. The resulting optimization problems are solved analytically by employing the technique of Lagrange multipliers. With the aid of some numerical examples, it is illustrated that the two schemes result in different jamming waveform design results, which is useful to guide jamming power allocation for various jamming tasks. It is also shown that the LPI performance of the smart jammer can be efficiently improved by exploiting the proposed jamming waveform design criteria.