We describe the atomic and molecular data that were used for the abundance analyses of FGK-type stars carried out within the Gaia-ESO Survey. We present an unprecedented effort to create a homogeneous line list, which was used by several abundance analysis groups to calculate synthetic spectra and equivalent widths. The atomic data are accompanied by quality indicators and detailed references to the sources. The atomic and molecular data are made publicly available in electronic form. In general experimental transition probabilities were preferred but theoretical values were also used. Astrophysical gf-values were avoided due to the model-dependence of such a procedure. For elements whose lines are significantly affected by hyperfine structure or isotopic splitting a concerted effort has been made to collate the necessary data for the individual line components. We also performed a detailed investigation of available data for line broadening due to collisions with neutral hydrogen atoms. Synthetic spectra calculated for the Sun and Arcturus were used to assess the blending properties of the lines. Among a subset of over 1300 lines of 35 elements in the wavelength ranges from 475 nm to 685 nm and from 850 nm to 895 nm we identified about 200 lines of 24 species which have accurate gf-values and are free of blends in the spectra of the Sun and Arcturus. For the broadening due to collisions with neutral hydrogen we recommend data based on Anstee-Barklem-OMara theory, where available, and to avoid lines of neutral species otherwise. Theoretical broadening data by R.L. Kurucz should be used for Sc II, Ti II, and Y II lines. For ionised rare-earth species the Unsold approximation with an enhancement factor of 1.5 for the line width can be used. Desirable improvements in atomic data were identified for a number of species, including Al I, S I, Cr II, Na I, Si I, Ca II, and Ni I.