Abstract. Global retrievals of near-infrared sun-induced chlorophyll fluorescence (SIF) have been achieved in the last few years by means of a number of space-borne atmospheric spectrometers. Here, we present a new retrieval method for medium spectral resolution instruments such as the Global Ozone Monitoring Experiment-2 (GOME-2) and the SCanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY). Building upon the previous work by Guanter et al. (2013) and Joiner et al. (2013), our approach provides a solution for the selection of the number of free parameters. In particular, a backward elimination algorithm is applied to optimize the number of coefficients to fit, which reduces also the retrieval noise and selects the number of state vector elements automatically. A sensitivity analysis with simulated spectra has been utilized to evaluate the performance of our retrieval approach. The method has also been applied to estimate SIF at 740 nm from real spectra from GOME-2 and for the first time, from SCIAMACHY. We find a good correspondence of the absolute SIF values and the spatial patterns from the two sensors, which suggests the robustness of the proposed retrieval method. In addition, we compare our results to existing SIF data sets, examine uncertainties and use our GOME-2 retrievals to show empirically the relatively low sensitivity of the SIF retrieval to cloud contamination.