To construct data warehouse with heterogeneous, self-governing and distributed database system is a challenge. Two problems have to be solved: one is how to collect data from heterogeneous data sources by an effective method, the other is how to clear up and change the format of the data collected from data sources. The opening and the interact-operation property of Web Services is Analyzed. A Data Collection System based on Web Services, which communicates with heterogeneous data sources with SOAP and HTTP, has been designed. A Data Collector is introduced to clear up and to transform data with composition technology. The pivotal XML technology of transforming and loading up the data are discussed.