Three different neutron activation analysis methods have been developed for the determination of ppb levels of iodine in food samples. The methods are based on the separation of iodine using (i) toluene extraction followed by bismuth sulfide coprecipitation; (ii) bismuth sulfide coprecipitation followed by radiochemical purification with palladium iodide; and (iii) radiochemical isolation by bismuth sulfide coprecipitation. The accuracy of these methods was evaluated by analysing replicate samples of reference materials. The measured values of iodine in A-11 Milk Powder, H-4 Animal Muscle and H-9 Mixed Human Diet from the International Atomic Energy Agency (IAEA) are statistically indistinguishable from the IAEA recommended values, and those for the Standard Reference Materials 1571 Orchard Leaves and 1577 Bovine Liver from the National Institute of Standards and Technology (NIST) are in good agreement with the NIST information values. The precision, in terms of relative standard deviation, is 5% at 50-100 ppb and 10% at 5-20 ppb levels of iodine. The absolute detection limits of these methods vary between 0.5 and 10 ng of iodine. All three methods were used to measure the iodine content of several food samples. The method involving bismuth sulfide coprecipitation followed by radiochemical purification provides the best detection limit and highest precision.