A 2‐year field experiment was conducted in Daejeon, South Korea, to investigate changes in the growth parameters of rice plants during the growth and model the rice yield loss as affected by the competition with Leptochloa fusca, Bidens frondosa and Ammannia coccinea under transplanted rice cultivation. Compared with B. frondosa and A. coccinea, L. fusca significantly reduced the number of tillers at the early growth stage, resulting in more reduction in the number of panicles and other yield components such as the number of spikelets, ripened grain ratio and the ultimate rice yield at the later growth stage. Among the components, the panicle number was the most negatively and positively correlated with weed density and rice yield, respectively. Cousens' rectangular hyperbola model predicted that weed competitiveness value represented by parameter β was 0.0337 and 0.0572 for L. fusca, 0.0124 and 0.0165 for B. frondosa and 0.0059 to 0.0061 for A. coccinea in 20.8 and 15.2 hill m−2 rice densities, respectively, indicating the competition effect of L. fusca on rice yield was greater than that of B. frondosa and A. coccinea. Weed densities corresponding to a 50% reduction in the rice yield were 27 and 18 for L. fusca, 81 and 61 for B. frondosa and 169–164 for A. coccinea in 20.8 and 15.2 hill m−2 rice densities, respectively. Weeds are more competitive in low‐density rice plots, leading to greater yield loss. In addition, the economic thresholds (ET) are positively correlated with rice densities and negatively correlated with weed competitiveness. Although low‐density rice transplanting may offer economic benefits during cultivation, optimal rice density and weed control strategies are important for securing a high yield. Therefore, the model and findings could be applied to decision‐making systems in rice farming where weed interference is present.