Abstract: This study aims to investigate a Cryptomeria (Cryptomeria japonica) plantation which was row-thinned with different thinning regimes at age 24 and under-planted with 4 species: Taiwan yellow false-cypress (Chamaecyparis obtusa var. formosana), Taiwan red false-cypress (Chamaecyparis formosensis), Luanta-fir (Cunninghamia konishii) and Taiwania (Taiwania cryptomerioides) in the thinned areas. After 21 years, the survival rates of under-planting trees revealed significant differences in different planting species (ranging from 59% in Taiwania to 31% in Luanta-fir), but not so in different thinning treatments. The DBH and height growths of the under-planting trees also showed significant inter-specific differences, however the effect of different thinning treatments was not significant. The DBH and height growth of Cryptomeria in the reserve strips were significantly higher than that in the control area. However, the volume in the control area was still larger than that in the thinned areas. We also observed a decrease of thinning effect on the periodic annual increments of diameter 13 years after thinning. The volume growth in the thinned areas obviously exceeded that in the control area 7 years after thinning. Besides, due to the serious wind damages occurring in the retained Cryptomeria trees 6 years after thinning, we recommend adopting broader retaining strips (20 m) in thinning operations to prevent wind damages. In the investigation of ground-layer vegetation, 97 plant species were recorded, with the Shannon index of diversity reaching 1.78 and the Simpson index of diversity reaching 0.66. The species richness was averagely lower in the retaining strips and control area than in the thinned areas, suggesting that different thinning treatments would create different environmental conditions in the understory and thus affect the composition and diversity of ground-layer vegetation. In the investigation of natural regeneration trees, 66 species and 2 010 trees (941 trees/hm2, averagely) were recorded, with the Shannon index of diversity reaching 2.92 and the Simpson index of diversity reaching 0.89. The diversity of natural regeneration trees in different thinning treatments revealed marginally significant differences, while the tree densities showed significant increases in the retaining strips and control area. These results suggest that the previous forest operations (thinning, under-planting, and tending etc.) would substantially affect natural regeneration in the under-story.