Abstract: To establish a scientific, suitable and accurate site quality evaluation system of natural forest, we used 123 analytic fir trees for simulating the optimal competitive index of the natural spruce-fir and broadleaf mixed stands among 6 competitive indexes. Eligible firs from 26 clear cutting plots and analytic trees were picked out to simulate the height growth curve of dominant trees and the height growth curve of worst trees. Then the suitable site index of competition (SIC) table was built by using the height growth curve of dominant trees and height growth curve of worst trees as guide curves. The optimal multiple regression prediction models for height growth were selected to analyze the height difference caused by competition. Then, the site quality of two site types was evaluated by using SIC. The results indicated: 1) The optimal competition index of natural spruce-fir and broadleaf mixed stands is CI1. Statistics of CI1 shows that different age grades have marked differences in competition. The pressure of competition decreases with aging. 2) Building competition tables with 2m class interval, a standard age of 60 years and 7 levels of site indexes is more reasonable. 3) The power exponent model of two site types is most precise in three multiple regression prediction models for height growth. 4) Evaluation of two site types using site indexes of the competition table shows that site type 1 belongs to 17m site index of competition, and site type 2 belongs to 15m site index of competition. The site type 1 is superior in site quality to site type 2.