Background: Multinucleon transfer reaction at low-energy collisions is considered to be a promising method for the production of new exotic nuclei, which are difficult to be produced by other methods. The theoretical studies are required to provide reliable predictions for the experiments and promote the understanding of the microscopic mechanism in multinucleon transfer reactions. Purpose: We provide a predictive approach for production cross sections, and testify how and to what extent the microscopic approach works well in multinucleon transfer reaction. Methods: We employ the approach TDHF+GEMINI, which combines the microscopic time-dependent Hartree-Fock (TDHF) with the state-of-art statistical model GEMINI++, to take into account both the multinucleon transfer dynamics and the secondary deexcitation process. The properties of primary products in multinucleon transfer process, such as transfer probabilities and primary cross sections, are extracted from TDHF dynamics by using the particle-number projection method. Production cross sections for secondary products are evaluated by using the statistical model GEMINI++.Results: We investigate the influence of colliding energies and deformation orientations of target and projectile nuclei on multinucleon transfer dynamics in the reaction 58 Ni+ 124 Sn. More nucleons are observed to transfer in the tip collision as compared to the side collision. The production cross sections for secondary fragments with TDHF+GEMINI calculations well reproduce the experimental measurements at energies close to the Coulomb barrier. At sub-barrier energy, the theoretical results gradually deviate from the experimental data as the increase of the number of transferred neutrons, implying the limitations of a single mean-field approximation in TDHF approach. Possible origins for this discrepancy are discussed. The total cross sections integrated over all the neutron pickup channels are in good agreement with the experimental data for all the energies. We compare the production cross sections of TDHF+GEMINI calculations with those from GRAZING model, and find that our approach gives a quantitatively good description as the semiclassical model, although there is no adjustable parameters for the reaction dynamics in the microscopic TDHF method. Conclusions: The microscopic approach TDHF+GEMINI reasonably reproduces the experimental data at energies close to the Coulomb barrier and well accounts for the multinucleon transfer mechanism. The present studies clearly reveal the applicability of TDHF+GEMINI method in multinucleon transfer reactions, which thus deserves as a promising tool to predict the properties of new reactions. * luguo@ucas.ac.cn hance the yield of exotic nuclei for an appropriate projectiletarget combination. To produce the new unstable isotopes experimentally, the optimal incident energy and projectile-target combination should be chosen to have the highest product cross section for the desired isotope. The reliable theoretical predictions are therefore required to guide ...