[Objective] The objective of the study was to elucidate the duplicate collinearity diagnosis method for soil nutrient and provide basis for establishment of multiple regression model of relationship between independent variable and depen- dent variable in fruit trees. [Method] An investigation on soil nutrients and ＇Fuji＇ fruit quality was conducted at 153 commercial apple orchards in the production re- gions of 51 counties in China from 2011 to 2012. At first, the regression equation was used to do the duplicate collinearity diagnosis, and then the partial least squares regression was used to analyze the weights of different soil nutrients that affected the fruit quality and screen out main soil nutrients. [Result] The results showed the conditional number between soil nutrients was 61.31, and the correlation coefficient of soil pH and available calcium, iron was 0.82 and -0.824 respectively, which demonstrated that there was multicollinearity between various soil nutrient contents. The soil organic matter, total nitrogen, available potassium, available nitro- gen, available phosphorus, available zinc, available iron and available boron showed positive effect Ioadings on fruit quality, while the }oadings and weights of the effects of available calcium and pH on fruit quality were negative. The variable importance for projection method was adopted to get the orders of the effects of major soil nu- trients on fruit quality, which was in the order of soil available calcium, nitrogen, available iron, available phosphorous, organic matter and available boron. The solu- ble solid content of the fruit was affected by soil organic matter, available phospho- rus, available iron, available boron and soil pH. [Conclusion] This study provided an analysis method and theoretical basis for the study and analysis on the relationships between dependent variables and multi-objective optimization, as well as the effects of soil nutrients on fruit quality in fruit tree.
