By acting as an ideal working fluid for the improvement of the heat transfer performance, nanofluids can enhance the process of convective heat transfer by adding nanoparticles to working fluids. As nanofluids can significantly improve heat exchange efficiency in the heat exchange systems, they have been studied extensively as a primary research topic in the past two decades. Owing to its good thermophysical properties, CuO is usually used as additive filler for improving the enhanced heat transfer performance of different base fluids. So far, various morphologies and structures of CuO nanoparticles and CuO composites (such as nanowires, nanoblocks, and nanospheres) have been synthesized. However, due to different preparation processes and systems, very few studies have explored the effect of CuO morphology on the enhanced heat transfer of nanofluids under the same conditions. In fact, the study of the influence of nanoparticle morphology and subsequent mechanism is significant for the promotion of its practical applications. Water and ethylene glycol have been the primary focus of traditional research works conducted on the base fluid of nanofluids. However, water and ethylene glycol have low boiling points; therefore, they cannot be exposed to higher temperatures. This limitation causes the devices (used for water and ethylene glycol) to become more prone to wear and corrosion, thereby limiting the wide applications of water and ethylene glycol in industrial production. As compared to water and ethylene glycol, dimethicone has the following advantages: Good thermal stability, non-toxicity, odorlessness, and recyclability. However, the thermal conductivity of dimethicone is too low, thereby leading to lower heat transfer efficiency. Therefore, researchers have tried to increase its thermal conductivity by employing various methods. Experimental studies have shown that the heat transfer performance of nanofluids can be enhanced by the addition of metal oxide nanoparticles having high thermal conductivity. In this study, we prepared CuO nanospheres, CuO nanoflowers, CuO nanorods, CuO nanowires-dimethicone nanofluids in order to verify the effect of CuO morphology on the thermal conductivity of nanofluids. The effects of particle volume fraction and shape on thermal conductivity were also studied. In addition, we carried out an in-depth analysis of its enhanced heat transfer mechanism and theoretical prediction model. The experimental results demonstrated that the structure and morphology of nanomaterials have a significant influence on the thermal conductivity of nanofluids. The increase of volume fraction indicates that there is a nearly linear relationship between thermal conductivity and volume fraction. At the volume fraction of 0.75%, thermal conductivities of CuO nanospheres, CuO nanoflowers, CuO nanorods, and CuO nanowires were 0.155, 0.168, 0.189, and 0.233 W/(m K), respectively. The corresponding thermal conductivities also increased by 7.00%, 16.11%, 51.62% and 60.78%, respectively. Through comparison, we conclude that the thermal conductivity of CuO nanofluids is positively correlated with the aspect ratio of CuO under the same conditions. Through a comprehensive comparison of experimental results and the classical theoretical prediction models, we found out that the Maxwell, Jeffrey, Bruggeman, and Looyenga models can meet the prediction requirements of thermal conductivity of spherical CuO nanofluids. The Hamilton-Crosser model can serve as a better predictor of the thermal conductivity of CuO nanofluids. In addition, we analyzed the heat transfer mechanism of CuO nanofluids. It was found that the higher aspect ratio of CuO nanorods and nanowires helps in increasing the probability of contact with each other. This, in turn, established an optimal condition for the hot carrier transport to achieve a minimum energy loss path, thus building a faster thermal conduction network.