This study was conducted to establish the elaborate nitrogen fertilization method to enhance N use efficiency in direct-seeded rice on flooded paddy. The nitrogen uptake by rice plants was insignificant until 25 days after seeding, and increased gradually thereafter. During this early growth stage, rice plants absorbed only the $4\%$ of basal applied N, while the $45\%$ of N fertilizer remained in the paddy soil. The absorption of basal N by rice plants was almost completed at 46 days after application. Nitrogen top-dressed at 5-leaf stage was well matched to crop nutrient demand, so it could be absorbed so actively in 8days after application. As a result, we could cut down the amount of N fertilizer to $36\%$ of the basal N level without significant difference in yield. Plant recoveries of fertilizer $^{15}N$ applied with different application timings were $7.8\%$ for basal, $9.4\%$ for 5-leaf stage, $17.1\%$ for tillering stage, and $23.4\%$ for panicle initiation stage, respectively. When urea was applied with nitrogen fertilization practice based on basal incorporation (BN), plant recovery of $^{15}N$ at harvest was $31.0\%$, which was originated from $13.7\%$ for grain, and $21.3\%$ of the fertilizer $^{15}N$ remained in the soil, and the rest could be uncounted. Plant recovery of fertilizer $^{15}N$ applied with nitrogen fertilization practice based on topdressing at 5-leaf stage (TN), where N rate was reduced by $18\%$ compared with BN, was $35.1\%$ (grain $15.6\%$), and $19.9\%$ of the fertilizer $^{15}N$ remained in the soil, and the rest could be uncounted. TN showed a higher $^{15}N$ recovery than BN because it was to apply N fertilizer at a time to well meet the demand of rice plant direct-seeded on flooded paddy. We concluded that TN would be the nitrogen fertilization method to enhance N use efficiency in direct-seeded rice on flooded paddy.