Abstract: Zinc slag is inevitable in the production of hot-dip galvanizing process for thin strip steel. The percentage of iron in zinc dross is about 5%. The increase of iron content in zinc slag makes it easier to produce bottom slag, which is deposited in the zinc pot and not easy to remove, and it will also bring great difficulties to the recycling of zinc resources in zinc slag. A vector auto-regressive algorithm was used to select three process variables, the temperature of zinc liquid, strip speed, and total aluminum content, to build a prediction model for iron slag content. The results show that it is feasible to apply the model to predict the iron slag content in the next 15 min. By adjusting the three process parameters, the generation of iron slag can be effectively reduced.