Consumers will be able to access applications and data anywhere in the world on demand by cloud computing which promises reliable services delivered through next-generation data centers. Cloud computing has a very broad application prospects such as virtualization, large-scale, dynamic configuration and many other characters. At the same time, there are many security risks such as privacy information leakage in the network by the rapid growing of network security threats. Security issues is the key issues constraining the development of cloud computing. The Privacy Protection Support Vector Machine (PPSVM) is widely concerned in secure multi-party computation (SMC). We propose a new optimized Privacy Protection Support Vector Machine classifier without Secure Multi-Party Computation for vertically partitioned data set which is not disclosing the private data. The novel approach is proved as being greater than traditional classification SVM on privacy-preserving by some experiments.