The dataset contains the results of airborne gravimetry realized by the GEOHALO flight mission over Italy in 2012. The intention was to show whether and how an efficient airborne gravity field determination is feasible in wide areas when using a fast jet aircraft like HALO at higher altitudes. Here, unlike in airborne gravimetry for exploration purposes, the aim is not primarily to reach the highest spatial resolution by flying as low and slowly as possible. A challenge for HALO would be to map areas (e.g., Antarctica) where only insufficient or no terrestrial gravity data are available to achieve a resolution which is better than that of satellite-only gravity field models. This is beneficial for the generation of global gravity field models which require a uniform, high spatial resolution for the gravity data over the entire Earth. The raw gravimetry recordings were recorded by the GFZ air-marine gravimeter Chekan-AM. Kinematic vertical accelerations were calculated from Doppler observations which were derived by GNSS carrier phase measurements (1 Hz). To remove the high-frequency noise, a low-pass filter with a cut-off wavelength of 200 s (corresponding to a half-wavelength resolution of approximately 12 km) was applied to both the Chekan-AM measurements and GNSS kinematic accelerations. To investigate how future airborne gravity campaigns using jet aircraft could be optimized, a dedicated flight track was repeated two times which shows that the equipment worked well also at higher altitude and speed. For the accuracy analysis 17 crossover points could be used. This analysis yielded a RMS of the gravity differences of 1.4 mGal which, according to the law of error propagation, implies an accuracy of a single measurement to be 1 mGal. The dataset is provided in as ASCII text (Lu-et-al_2017-001_Tracks_GEOHALO.txt) and is described in the README. For a detailed description of the set-up and analysis of the data, please see Biao et al. (2017, http://doi.org/10.1002/2017JB014425).