While many researches of Host Anomaly Detection System using system calls under UNIX/UNIX-like systems have been done but little in Windows systems, we do the similar research under Windows platforms via tracing the sequences of Windows Native APIs which are considered as the Windows system calls. In this article, we first introduce Native API briefly and then divide the captured sequences with slide window method to establish normal pattern database. Then Support Vector Machine Method is used for anomaly detection due to its advantages in small-scale dataset and generalization capability. The main purpose of this paper is to prove that Windows Native APIs are plausibly possible data source for Host Anomaly Detection System under Windows platforms.