In this paper, a formula is proposed to express the resolution evaluation coefficient (REC) in ghost imaging (GI). In discussing the relationship between REC and measurement conditions, a second-order correlation function across multiple measurement conditions is presented, on which a parametric differential ghost imaging (PDGI) is proposed to improve the resolution of GI. Theoretical analysis proves that more detailed information about objects can be revealed by increasing the number of measurement conditions. Thus, point spread functions (PSFs) of second-order correlation functions are employed to narrow the shape of PSFs. Experimental results confirm the effectiveness of the proposed method.