Abstract: To deal with the problem of radar emitter recognition caused by noise environment, this paper presents a new method for emitter recognition based on backward cloud model and attribute similarity. In this method, a radar emitter database including noise data according with the reality is first constructed, the cloud numerical characteristic is then calculated based on backward cloud model. After that, the method for determining recognition weight of coefficients and a new classification implement based on backward cloud model and attribute similarity are proposed. Simulation results show that the proposed method can deal with the randomness and vagueness caused by noise environment much better and can conduct emitter recognition effectively in the adverse noise environment.