Compressor stations on natural gas pipelines are typically composed of multiple parallel compressors for operational flexibility. The aim of load sharing optimization is to operate the compressor units of a gas compression station in an energy efficient way while satisfying the varying gas demand. This paper presents a problem formulation for compressor load sharing optimization, as well as a novel method to track the performance characteristics of gas compressors using thermodynamic models and historical operating data. An implementation of the proposed algorithm together with a dedicated case study concerning a station with 10 gas turbine driven compressor units are presented. The optimization results, which are based on simulations with actual operating data spanning over a 1 year duration, indicate an annual fuel saving potential of over 5% potentially leading to improved profitability and to a significant reduction of CO2 emissions from the gas turbine drivers. (C) 2016 Elsevier Ltd. All rights reserved.