Abstract: Classification is important in data mining.Many classifying algorithms often deal with some discrete data，so the paper puts forward a new classifying algorithm based on compound particle swarm optimization，it can classify the data with the continuous and discrete attribute.In order to advance the classification validity and the efficiency，the paper codes with the compound particle structure，and gets the discrete data by dispersing the continuous attribute data，then classifies them by the particle swarm optimization，which translate the problem of the mix-data classification into the combination optimize problem with 0-1.The experiment results prove that the algorithm has a good effect and speediness.