Atomic line opacities play a crucial role in stellar astrophysics. They strongly modify the radiative transfer in stars, therefore impacting their physical structure. Ultimately, most of our knowledge of stellar population systems (stars, clusters, galaxies, etc.) relies on the accuracy with which we understand and reproduce the stellar spectra. With such a wide impact on Astronomy, it would be ideal to have access to a complete, accurate and precise list of atomic transitions. This, unfortunately, is not the case. Few atomic transitions had their parameters actually measured in the laboratory, and for most of the lines the parameters were calculated with low precision atomic energy levels. Only a small fraction of the lines were calibrated empirically. For the purpose of computing a stellar spectral grid with a complete coverage of spectral types and luminosity classes, this situation is rather limiting. We have implemented an innovative method to perform a robust calibration of atomic line lists used by spectral synthesis codes called ALiCCE: Atomic Lines Calibration using the Cross-Entropy algorithm. Here we describe the implementation and validation of the method, using synthetic spectra which simulates the signal-to-noise, spectral resolution and rotational velocities typical of high quality observed spectra. We conclude that the method is efficient for calibrating atomic line lists.