Few thermal conductivity models contain the effect of particle shape and the measured value is significantly larger than the conventional prediction when containing specially shaped particles are two challenges for nanofluids. This paper, inspired from the effect of interfacial layer, proposes a thermal conductivity model by the heat conduction analysis in arbitrary direction of a finite length cylindrical nanoparticle in a fluid medium. In the model, the effective thermal conductivities of nanofluid in axial and radial direction are deduced respectively and then equalized as the two principal axes of an equivalent anisotropic material. Then thermal conductivity coefficient transformation of anisotropic material on arbitrary direction is implemented in the model building process. Comparing with other classical models, the proposed model takes into account the particle shape viz. radius and length of the cylindrical particle. The theoretical predictions on the enhanced thermal conductivity agree quite well with the available experimental data since 95% of the relative errors are within 10%. (C) 2016 Elsevier B.V. All rights reserved.