In order to realize the targets of animal health culture, which including animal health, human being health and environment health, early warning architecture contents of animal health culture was put forward. Taking the pork as an example, pork price early warning system model was established based on BP neural networks and the system platform was designed and implemented by combined programming of C# and MATLAB. Early warning system of chicken products quality and breeding disease was implemented by using Java EE architecture in taking the broiler as an example. Early warning system of natural environment was designed with C# language according to both animal manure excretion and theoretical stocking rate of natural grassland in a certain area. Early warning system of environment parameters in intensive culture was implemented by using of embedded program design. When the fluctuation range of pork price exceeds 15%, daily mortality of broilers and complaint rate of chicken products exceed prescribed threshold, breeding number of animal is more than theoretic capacity of natural environment, breeding environment parameters exceed the set range, the early-warning messages would be sent. Early-warning of animal price and capacity of natural environment can provide the basis for government administrators and guide animal production, the economic loss and environmental damage would be avoidable by reducing price fluctuation and excessive breeding. Early-warning of animal products quality, epidemic and environment parameters of animal culture, from all the aspects of production management, can ensure the safety of animal products and also improve the animal welfare.