To detect the gross errors efficiently in the process of data reconciliation, a new target function is proposed based on robust M-estimators. Robustness of this new target function is evaluated through the influence function. Then, the performance of the new target function is illustrated through a linear data reconciliation example and a nonlinear one. Simulation results show that the new target function gives promising results for data reconciliation and gross error detection. Besides, the proposed new target function has an advantage over the Cauchy and Huber functions on detecting small gross error.