To assess the radiation use efficiency (RUE) of different rice cultivars under transplanting (TP) and direct-sowing (DS) systems and scrutiny of weed effect, a split-plot factorial experiment was conducted based on a randomized complete block design with three replications in 2011 at Haraz Extension and Technology Development Center in Mahmoudabad city, Mazandaran. The main plots were two planting systems (TP and DS) and subplots were factorial combinations of three rice cultivars under weed infested and weed free conditions. The results showed that RUE and grain yield of rice cultivars significantly reduced in weed infested plots especially in DS system. The average RUE of line 843, Khazar and Tarom in weed-free conditions were 3.18, 3.22 and 2.69 g MJ^-1, respectively, while it reduced to 3.08, 2.57 and 2.35 g MJ^-1, respectively, under weedy conditions. The differences of yields were mainly due to differences in biomass and harvest index, and were not associated with RUE. Light extinction coefficient (K) of line 843, Khazar and Tarom were 0.44, 0.45 and 0.71, respectively. Under high weed interference (as seen under DS method), Tarom could better suppress weeds and had lower yield loss due to its taller plant height and greater K. Under no or low weed interference (as seen under TP method), line 843 was ranked higher in terms of final yield due to its smaller K, greater leaf area and higher harvest index.