Quick and non-destructive methods for estimating leaf area were important for future related research of Ixora. In order to estimate leaf area of two varieties of Ixora, mature leaves of I. coccinea var. coccinea and I. coccinea var. bandhuca were collected as the experiment materials. The WinFolia software was used to measure leaf shape indices such as leaf area(LA), leaf perimeter(LP), leaf vertical length(VL), leaf blade length(L), leaf horizontal width(W), LW, LL, WW, and L/W. The respective regression analysis between leaf area and other leaf shape indices of the varieties of Ixora were carried out by using 8 common regression models. Besides, the general regression equation that suitable for two Ixora varieties were established. The results indicated that the correlation and the multiple correlation coefficients between the leaf area and the leaf shape indices，such as LW, LP, VL, L, W, LL, and WW, were at extremely significant level (P<0.01); the correlation and the multiple correlation coefficients between the leaf area and LW were the highest, respectively; leaf area could be predicted better with LW by quadratic function, composite function and power function; among the three kinds of regression equations that both suitable for two Ixora varieties, the composite function and power function could be better to predict leaf area.