Abstract: A modified SICC (serial identification with collective compensation) method used in gross error detection and data reconciliation for chemical process is introduced in this paper. By using temporal redundancy, upper and lower bounds of process measurements are added in the modified SICC method to avoid the misidentification of gross errors. To prevent the computation of project matrix, the measurements are compensated by using the estimated gross error magnitudes after each gross error is identified, and after that, MT (measu rement test) method is used to find the gross error candidates. To avoid the singularities from appearing after gross error compensation, necessary cycle detection is added. Simulation results on a steam metering system verify the effective ness of the modified algorithm.