Fracturing technology is an indispensable part in the development of low permeability reservoirs. In order to solve the problems of water consumption and formation damage, carbon dioxide (CO2) fracturing is treated as an alternative formation transformation and production enhancement technology for water-based fracturing. Several researches on fracturing mechanism and advantages of both liquid carbon dioxide (L-CO2) and supercritical carbon dioxide (Sc-CO2) mainly focused on the comparison and analysis of fracturing effects between different fracturing fluids, without evaluating the influence of fluid injection rate and core length for the effects of fracturing. Hence, as the important parameters in fracturing scheme design, it is necessary to study the effect of injection rate on L-CO2 fracturing and Sc-CO2 fracturing, as well as the effect of core length on L-CO2 fracturing. In this study, fracturing experiments with coreholder system, CT scanning, permeability and porosity tests are carried out to investigate and analysis breakdown pressure, permeability, porosity, 2D and 3D reconstruction images of each post-fractured tight sandstone cores. Since the micro-fractures formed by fracturing in the core are not easily detectable with limited experimental precision, the post-fractured core spontaneous imbibition experiments are introduced to reflect the generated fracture properties and the existence of secondary micro-cracks inside samples. Combined with one mathematical model of spontaneous imbibition oil recovery, a complete fracturing evaluation system under laboratory conditions is established on the basis of post-fractured core breakdown pressure, permeability and porosity, induced fracture morphology and propagation, spontaneous imbibition weight, imbibition potential and imbibition oil recovery, to evaluate fracturing performance in four major objectives: fracturing with different fracturing fluids, L-CO2 fracturing with different injection rate, Sc-CO2 fracturing with different injection rate, and L-CO2 fracturing with different core length. The results show that water-based fracturing tends to generate single and smooth fracture inside core, which has the highest breakdown pressure and lowest permeability. L-CO2 fracturing tends to form the short and narrow main fracture with transverse micro-cracks inside core. Sc-CO2 fracturing has a trend in creating the fracture network inside core, which has lowest breakdown pressure and highest permeability and calculated imbibition oil recovery. Besides, high-injection-rate L-CO2 fracturing path is specially long and straight with many longitudinal micro-fractures, resulting in post-fractured core with relatively high breakdown pressure, larger permeability, greater imbibition efficiency and oil recovery. In addition, with the increase of injection rate in Sc-CO2 fracturing, the fracture network structure created in the core becomes less distinct. Low-injection-rate Sc-CO2 fracturing has the best effect on core transformation because of post-fractured core with higher permeability, greater imbibition efficiency and oil recovery. Furthermore, L-CO2 fracturing is not ideal for transformation of overly long core sample.