The level of genetic variation within a breeding population affects the effectiveness of selection strategies for genetic improvement. The relationship between genetic variation level within Pinus tabuliformis breeding populations and selection strategies or selection effectiveness is not fully investigated. Here, we compared the selection effectiveness of combined and individual direct selection strategies using half- and full-sib families produced from advanced-generation P. tabuliformis seed orchard as our test populations. Our results revealed that, within half-sib families, average diameter at breast height (DBH), tree height, and volume growth of superior individuals selected by the direct selection strategy were higher by 7.72%, 18.56%, and 31.01%, respectively, than those selected by the combined selection strategy. Furthermore, significant differences (P < ​0.01) were observed between the two strategies in terms of the expected genetic gains for average tree height and volume. In contrast, within full-sib families, the differences in tree average DBH, height, and volume between the two selection strategies were relatively minor with increase of 0.17%, 2.73%, and 2.21%, respectively, and no significant differences were found in the average expected genetic gains for the studied traits. Half-sib families exhibited greater phenotypic and genetic variation, resulting in improved selection efficiency with the direct selection strategy but also introduced a level of inbreeding risk. Based on genetic distance estimates using molecular markers, our comparative seed orchard design analysis showed that the Improved Adaptive Genetic Programming Algorithm (IAPGA) reduced the average inbreeding coefficient by 14.36% and 14.73% compared to sequential and random designs, respectively. In conclusion, the combination of the direct selection strategy with IAPGA seed orchard design aimed at minimizing inbreeding offered an efficient approach for establishing advanced-generation P. tabuliformis seed orchards.