Time reversal technique has the adaptive time-space focusing characteristics, which has been widely used in communication systems, imaging systems, and power combining systems. However, the ideal time reversal processing cannot be implemented in an actual imaging system and some diffusion phenomenon has been observed. In this paper, the diffusion phenomenon of the time reversal field in an imaging system is analyzed based on the time reversal cavity theory. Since the corresponding absorption source cannot be set in an imaging process, the time reversal field will continue to disperse after the convergence. Therefore, the field produced by the time reversal cavity will be similar to the sinc-function near the source. The diffusion field will result in mutual interference between the imaging targets. In a traditional time reversal multi-target imaging system, weaker targets can easily be concealed and artifacts may occur. In this paper, a multi-target imaging technique based on the elimination of the time reversal field diffusion is proposed. In order to eliminate the effect of the diffusion field, the Clean algorithm is used. The Clean algorithm is a de-convolution algorithm, which can effectively suppress the side lobe signal. By using the Clean algorithm in the time reversal imaging system, the interaction between multi-targets can be eliminated. Full-wave simulation shows a good performance of the proposed method. In practice, the time reversal mirrors are used to replace the time reversal cavity, for the fully closed time reversal cavity cannot be implemented. The effects of the time reversal mirrors have also been analyzed in this paper. The result shows that the positions of the time reversal mirrors have an significant influence on the reversed field distribution, which affects the Clean algorithm and the proposed imaging method. In order to eliminate the influence of time-reversal mirror position, an effective time reversal signal equalization algorithm is proposed. In the equalization algorithm, the amplitude of the time reversal signal in the time reversal mirrors is adjusted according to both the distance and the intensity. The proposed equalization algorithm can keep the time reversal field stable and provide effective support for the imaging method.