A fast and efficient algorithm-Recursive Measurment Error Estimation Identification (RMEEI) method for bad data analysis is further developed in this paper. By using a set of Linear recursive formulae, state variables, residvals and their variances are updated after the removal of a measurment from suspected data set to the remsining dat set (or in the reverse direction). Neither a re-estimatiion nor residual sensitivity matrix in meeded in the identification process, which increases the computational spead greatly. Digital tests have been done to compare the RMEEI methods in identification performance and comoputational speed. The real-time operation experience of the RMEEI method in EMS of the North East China system control center are also given.