With presence of hybrid bad telemetry data and error parameter in power system, the validity of parameter identification and estimation methods of whole network cannot be guaranteed due to the fact that bad data will affect the parameter identification accuracy. It presents a detection and identification approach of bad-data based on augmented state estimation. First of all, the bad data are estimated whether they are bad telemetry data or parameters with error according to the residual balance degree. After deleting bad telemetry data, parameters with errors are kept within a certain area using node partition and then are modified according to augmented state estimation. The example results show that the proposed method can identify the bad telemetry data and parameters with error effectively, and interaction between the bad data can be avoid through parameter partition, so that the estimation accuracy of the suspicious parameters can be improved.