Abstract: According to genetic model for triploid endosperms with interaction of genotype×environment proposed by Zhu Jun(1994, 1996), it had been proved by Monte Carlo Simulation that minumum norm quadratic unbiased estimation is feasible for quantitative genetic analysis of data from unequal experiment design. The results of Monte Carlo simulation indicated that the biases or power values of estimated genetic parameters did not exist significant difference between equal and unequal experiment design.Therefore, the unbalance data from unequal experiment designs could also be used as genetic analysis for estimating genetic variance components and genetic covariance components, and predicting genetic effect values.