The aim of this work was to assess the influence of various formulation parameters on the incorporation of a poorly water-soluble crystalline drug into nanoparticles. For this purpose, the influence of the polymer (polylactic acid, polysebacic acid terminated with lithocholic acid, and polysebacic acid-co-lithocholic acid) as well as the effect of the dispersion medium (aqueous phases at different temperatures, saline medium and ethanol) on the encapsulation was investigated. 3H-labelled drug was used in order to determine the loading efficiency by liquid scintillation counting. The solubility of the drug in the various polymer materials was assessed by differential scanning calorimetry (DSC). The solubility of the drug in the different dispersion media was then determined by gas chromatographic-mass spectrometric measurements. The highest loading ratios were obtained using poly (lactic acid) (PLA). However, the drug solubility in the polymers, determined by DSC analysis, cannot be considered as predictive for encapsulation efficiency. The study of the influence of the liquid outer phase showed that the encapsulation efficiency increased when the drug solubility in the dispersion medium (before acetone evaporation) decreased. These experiments made it possible to propose a mechanism to account for the leakage of the crystalline drug during the nanoprecipitation process. So, when acetone is eliminated by evaporation, the drug solubility in the dispersion medium decreases, leading to the formation of crystals. During nanoparticles storage, the crystals continue to grow, the nanoparticles serving as drug reservoirs. These findings highlight the importance of using a polymer with a specific affinity for the drug, and a dispersion medium with the lowest drug solubility to achieve an efficient encapsulation of a crystalline drug.