To compare the breath-hold, respiratory-triggered, and free-breathing techniques in diffusion-weighted magnetic resonance imaging (MRI) for the evaluation of focal liver lesions on a 3.0T system.One hundred and fifty-two patients with focal liver lesions underwent routine MR and breath-hold, respiratory-triggered, and free-breathing diffusion-weighted imaging (DWI)^[1,2,3] of the liver on a MR 3.0T system, 75 of them were examined by the strategy (1): All DWI were performed with identical parameters except for signal averages (two for respiratory-triggered, one for breath-hold, and four for free-breathing). The sweep range included the whole liver, and the b values were 100, 800 s/mm²,respectively.Each DWI’s scanning time were 20s for breath-hold, 2min for respiratory-triggered, and 1 min for free-breathing. Another 77 patients were examined by the DWI strategy (2): The scanning parameters of the three techniques were just the same, and excitation numbers were all changed into 2，only scanned 10 layers, just the local lesions were examined. Overall scan times were controlled in 20s for breath-hold, 25s for free-breathing, 1 min for respiratory-triggered DWI. The b value were still 100, 800 s/mm². 2 doctors assessed and measured the data respectively.In the DWI strategy (1), for all types of breathing acquisitions, the focal liver lesions’ detection and accuracy rates had no statistical differences（P>0.05）. The focal lesions’ ADC values in three groups showed high correlation (when b=100 and 800, the ICC were 0.970 and 0.887, respectively), but the normal liver parenchyma’s ADC values of the three kinds DWI did not match when b = 800（P<0.0167）, the consistency of performance was moderate（ICC=0.425）, while the ADC values had no statistical difference when b=100（P>0.0167）, and they showed a strong consistency（ICC=0.646）. But the CNR(contrast to noise ratio), SNR(signal to noise ratio) of breath-hold DWI was lower than respiratory-trigger and free-breathing DWI（P<0.0167）. In the DWI strategy (2), the three DWI acquisitions’ lesion detection (90.1 ~ 95.6%) and accuracy rate (85.4 ~ 89.7%)had no statistical difference (P > 0.05), and the three groups’ ADC values of normal liver (b = 100 and 800, the ICC was 0.701 and 0.517, respectively) and lesions (b = 100 and 800, the ICC was 0.947 and 0.765, respectively) showed high correlation, but the free-breathing DWI’s CNR, SNR was significantly lower than the respiratory-triggered and breath-hold DWI (P < 0.0167). More detail ADC value and ROC curve referent to Figure.1-2 for strategy (1), Figure.3-4 for strategy (2). Specific SNR values the two strategy were listed in Table.1.For the patients who want to do the whole liver DWI examination, the free-breathing DWI was best, as it can achieve better image quality in a short time; but for the patients with focal liver lesions who need to do the DWI scan to evaluate the local lesions or the responds of medical treatment, the breath-hold DWI was optimal. Generally speaking, the respiratory-triggered DWI of the 3T MRI system had higher SNR and CNR in any case, and it was the best DWI acquisition technique.