Abstract: The objective of this study was to analyze the simple correlation coefficients in order to understand the pair-relations and carry out a path analysis to determine direct effects. Character associations between yield components can be used as the best guide for successful yield improvement by indirect selection. Thirty-six safflower accessions were evaluated for phenotypic traits at 4 locations namely Katumani, Kinamba, Lanet and Naivasha for two long rain seasons in two years. The experimental design was a Partially Balanced Lattice design with 3 replications. Path analysis of yield components revealed that components with the highest correlation to yield also had the highest direct effect to yield i.e., primary branches, capitula/plant, effective capitula, seeds/capitula and 100 seed weight. Two approaches can be suggested in selecting genotypes for specific traits to improve seed yield in safflower: Select genotypes with high number of seeds/capitulum, which is directly responsible for seed weight of capitulum as is evident from the high positive correlation between the two. Alternatively select genotypes with high 100 seed weight as this has a high significant (p ≤ 0.05) positive association with seed yield. Generally simple correlations differed at the same location every year. The high correlation and direct effects of number of capitula/plant, effective capitula/plant and seeds/capitulum on seed yield suggests selection for these components will improve seed yield in safflower.