Previous studies have revealed considerable Cd isotope fractionations in seawater, which can be used to study the marine cycling of this micronutrient element. The low Cd concentrations that are commonly encountered in nutrient-depleted surface seawater, however, pose a particular challenge for precise Cd stable isotope analyses. In this study, we have developed a new procedure for Cd isotope analyses of seawater, which is suitable for samples as large as 20 L and Cd concentrations as low as 1 pmol/L. The procedure involves the use of a (111)Cd-(113)Cd double spike, co-precipitation of Cd from seawater using Al(OH)(3), and subsequent Cd purification by column chromatography. To save time, seawater samples with higher Cd contents can be processed without co-precipitation. The Cd isotope analyses are carried out by multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS). The performance of this technique was verified by analyzing multiple aliquots of a large seawater sample that was collected from the English Channel, the SAFe D1 seawater reference material, and several samples from the GEOTRACES Atlantic intercalibration exercise. The overall Cd yield of the procedure is consistently better than 85% and the methodology can routinely provide ε (114/110)Cd data with a precision of about ±0.5 ε (2sd, standard deviation) when at least 20-30 ng of natural Cd is available for analysis. However, even seawater samples with Cd contents of only 1-3 ng can be analyzed with a reproducibility of about ±3 to ±5 ε. A number of experiments were furthermore conducted to verify that the isotopic results are accurate to within the quoted uncertainty.