【Objective】To explore the feasibility and simplicity of intelligent automatic registration ultrasound-CT/ MR fusion imaging based on liver surface in localization of focal liver lesions. 【Methods】 Thirty patients with detected focal liver lesions by contrast- enhanced CT or MR were enrolled for ultrasound- CT/MR fusion imaging using the PercuNav fusion imaging system in PHILPS EPIQ7. Both intelligent automatic registration ultrasound- CT/MR fusion imaging based on liver surface（intelligent method）and the conventional internal plane method（manual method）were used for ultrasound-CT/MR fusion imaging. The success rate of registration，the initial registration error and the times of fine-tuning were compared between these two methods.【Results】In all 30 patients，the success rates of registration were both 96.67%（29/30）using intelligent method and manual method. There was no significant difference between these two methods when compared the initial registration error and the times of fine- tuning （P>0.05）. According to the further stratified analysis，in 10 lesions in the left liver，the initial registration error of the manual method was less than that of intelligent method，the difference was statistically significant（P=0.00）. Although the times of fine-tuning of the manual method was less than that of intelligent method，there was no statistically significant difference（P=0.09）；In 20 lesions in the right liver，the initial registration error and the times of fine-tuning of the intelligent method were superior to those of the manual method. The differences were statistically significant （P<0.05）. 【Conclusion】 Intelligent automatic registration ultrasound- CT/MR fusion imaging based on liver surface is a feasible method with high success rate for ultrasound-CT/MR fusion imaging. Compared with the conventional internal plane method，the fusion imaging process is more simple and efficient for the lesions in right liver. It helps to reduce experience dependence of fusion imaging for the operators.