Managing production of complex oil fields with multiple wells and coupled constraints remains a challenge for oil and gas operators. some technical works developed strategies for integrated production optimization to assist production engineers in reaching best operating conditions. however, these works have neglected the uncertainties in the well inflow-performance curves and production processes, which may have a significant impact on the operating practices. the uncertainties may be attributed to measurement errors, oscillating behavior, and model inaccuracy, among others. to this end, this work proposes a formulation in mathematical programming for robust production optimization of gas-lifted oil fields. the formulation represents system-measured and simulated sample curves that reflect the underlying uncertainties of the production system. this representation leads to a robust milp formulation obtained from piecewise-linear approximation of the production functions. further, this work presents results from a computational analysis of the application of the robust and nominal formulations to representative oil fields available in simulation software.