采用回归最优设计方法，研究了杂交晚粳甬优538作单季直播时播种期（X1）、播种量（X2）、尿素用量（X3）和各期施氮比例（X4）对产量等8个主要性状的作用效应。结果表明，栽培效应播种期＞播种量＞施氮比＞尿素用量；播种期对齐穂期、结实率、产量等6个性状，播种量对最高苗数、每穂总粒数、产量等7个性状，尿素用量对分蘖数、有效穂数等4个性状，施氮比对株高、千粒重、产量等6个性状有显著影响，且因子间存在一定的互作效应。全因子模拟筛选单产≥750 kg/667 m2的优化措施：播种期6月2-5日，播种量1.25~1.41 kg/667 m2，施氮量（尿素）36.3~40.0 kg/667 m2，其中幼苗肥占28%~32%、分蘖肥占32%~28%、壮秆肥占18%~22%、穂肥（促花）占22%~18%。

Using the optimum regression design, the effects of seeding date （X1）, seeding rate （X2）, urea rate （X3） and nitrogen proportion in various growing stages （X4） were studied on 8 agronomic traits of japonica late hybrid rice Yongyou 538 by direct seeding. The results showed that the order of the cultivation effects was X1＞X2＞X4＞X3. There were various significant effects of seeding date to 6 agronomic traits （full heading date, seed setting rate, yield, etc）, seeding rate to 7 traits （maximum seedlings, total grains per panicle, yield, etc）, urea rate to 4 traits （tillerings, effective panicles, etc） and nitrogen proportion to 6 traits （plant height, 1000-grain weight, yield, etc）, and had interactions among factors. The analysis by the full factorial simulation screen indicated that in order to obtain higher rice yield ≥ 750 kg/667 m2, the optimal cultivation requirements should be adopted as follows: seeding date June 2-5, seeding rate 1.25~1.41 kg/667 m2, amount of urea 36.3~40.0 kg/667 m2, and nitrogen proportion 28%~32% in seedling stage, 32%~28% in tillering stage, 18%~22% in booting stage, and 22%~18% in flowering stage.