Fault-filling calcite crystals sampled from the Eocene/Oligocene Gondrecourt graben, Paris Basin, Eastern France, have been studied in order to test the potentiality of calcite (U-Th)/He dating, based on recognized He retention behavior in crystal lattice at surface temperature (Copeland et al., 2007). The samples have been selected because of their relatively old Eocene to Oligocene precipitation age and cold thermal history (<40 °C since precipitation). They were sorted into three main tectonic and morphological groups in order of precipitation, including (1) micro-fracture calcites, (2) breccia and associated geodic calcites, and (3) vein and associated geodic calcites. (U-Th)/He dating of 63 calcite fragments yields ages dispersed from 0.2 ± 0.02 to 35.8 ± 2.7 Ma, as well as two older dates of 117 ± 10 and 205 ± 28 Ma (1σ). These He ages correlate with grain chemistry, such as Sr, ΣREE concentrations or (La/Yb)N ratios, likely reflecting parent fluid evolution. Only the oldest He ages, which correspond to the most recently precipitated crystals, have preserved the total 4He budget since precipitation. To better understand both the age dispersion and why calcites precipitated earlier show younger ages, 4He diffusion experiments have been conducted on 10 Gondrecourt calcite fragments from 3 samples with He ages of ∼0.2-6 Ma. Additionally, a crystallographic investigation by X-ray diffraction (XRD) performed on similar samples reveals that crystal structure evolves with increasing temperature, beginning with micro-cracks and cleavage opening. These XRD results shed light on the (U-Th)/He data, indicating that, in fault-filling calcite, He retention is controlled by multiple diffusion domains (MDD) with various sizes, and therefore, evolves through time with strong consequences on (U-Th)/He age. We thus interpret the Gondrecourt calcite (U-Th)/He age scatter as a consequence of the production of defects due to successive calcite crystallization phases associated with the deformation history.