Provable data possession is an important research field in cloud storage security.It allows the data owners remotely checking the integrity of their outsourced data without downloading all files.There have been many batch PDP schemes,but most of them did not consider the error location after the data of users were corrupted.A few batch PDP protocols can identify only the servers in which the corrupted data stored or the clients to which the corrupted data belongs.This study puts forward a method which utilizes location tags to help the third party auditor locating the error data quickly.Based on work by Zhou et al.,an error locating batch provable data possession scheme is proposed in multi-user and multi-cloud setting by using Merkle Hash tree to create data location tags.The proposed protocol can quickly locate the corrupted data owners and the servers where the error data stored after the batch verification fails.The proposed scheme is provably secure in random oracle model,and the performance analysis shows that the scheme has higher error locating ability and efficiency than other schemes that only have single location function.