Field experiment was conducted out Al-Anbar in winter season of 2011/2012 to investigate decision making of selection to the genotypes of caraway Carum carvi L. thus, to introduce in consequent breeding program and test effectiveness of GT Biplot and two model of multi-trait selection, master and against checks jointly in GGE Biplot software for seven genotypes by six traits of caraway. Genotypes were randomly distributed underwith RCBD design with three replicates. Results revealed that these techniques were effective to derived information from data. Where, GT Biplot extracted 92.6% of total variation. Furthermore, it was indicated that was significant association between plant height and umbellets no. which those traits were the closest to yield. Also, branches no. significantly correlated with umbels no. the previous technique distributed the genotypes selectedF1 and Balady with yield per plant and proportionally with plant height and umbellets no., Mosul and Bery with branches and umbels no. multi-trait selection-master derived five genotypes from seven viz., Bery, Balady, Mosul, Sury and rather selectedF1. Where, Bery is selected depended on umbels diameter and umbels no., Mosul on branches no., Balady on yield per plant and umbellets no. and Sury on plant height. Whereas, second model-against checks is represented as a filter for these traits and complementary for first model-master which umbel diameter and umbel no. were passed for genotype Bery, plant height and umbellets in Balady and plant height in Mosul. Finally, it could be inferred from these three techniques that the best three genotypes were namely Bery, Balady and Mosul. Moreover, the second model was rather similar to GT Biplot. Generally, those techniques were effectively mining information from multi-traits. Therefore, it could be recommended to introduce these genotypes by multi-traits in different years by locations trials to obtain most precisely informative data in order to establish breeding program and improvement of caraway.