Coffee is the world's most widely traded tropical agricultural commodity next to oil and is the back bone to the Ethiopian economy. The associations between yield and its components offer important information for plant breeders. Hence, this experiment was conducted at Agaro research station, southwest Ethiopia, to determine the extent of correlations among yield and its related traits of 49 coffee accessions and to assess the direct and indirect effects of these traits on coffee yield. The experiment was conducted in simple lattice design with two replications. Data on yield and its components were recorded and analysis of variance, correlation and path coefficient analysis were performed. The result revealed that significant (P< 0.05) differences among the coffee accessions for all traits except for percent of bearing primary branches, leaf area, bean thickness and rust incidence. All traits showed non significant correlation with coffee yield at phenotypic level. However, morphological traits, namely, average inter node length on stem, stem diameter, angle of primary branches, fruit length, fruit width and fruit thickness exhibited positive and significant (P< 0.05) correlation with coffee yield at genotypic level and are important components to improve it. Besides, path coefficient analysis showed that average-inter nodes length on main stem, stem diameter, angle of primary branches, fruit length and fruit thickness having positive direct effects, exhibited positive significant genotypic correlation with coffee yield and these traits are important for direct selection to improve coffee yield. On the other hand, fruit width showed negative direct effect but positive and significant genotypic correlation with coffee yield and can be used for indirect selection. In conclusion, average inter node length of main stem, stem diameter, angle of primary branches, fruit length, fruit thickness and fruit width can be used as direct and indirect selection criteria to improve coffee yield.