Lentil is an important legume crop globally with an annual production of around 6.3 million tons. Pakistan stands at the 49th position producing 4668 tons of lentil from 7428 hectares with an average yield of 570 kg/ha. A lack of high-yielding varieties is one of the major reasons for low yield, resulting in an approx. 31% decrease in the cultivation area. In the present study, 649 accessions of lentil representing Pakistan, USA, and Syria were studied for yield and yield-contributing traits for three consecutive years. Accession 5930 performed best in all three years, having a seed yield (SY) of 192.84 ± 9.05 g/m and a biological yield (BY) of 534.20 ± 25.79 g/m. Overall, SY has a significant positive association with BY, pods per plant (NP), pod weight (PW), harvest index (HI), and plant height (PH). PCA, heritability, and genetic advance also suggested these traits as effective selection indicators. A K-mean cluster analysis based on Wilks lambda highlighted that accessions with a higher SY, BY, NP, PW, and NB were grouped in Clusters III, V, and II during the first, second, and third years, respectively. During all three years, genotypes in the HI range 10.1-15% had the highest biological yield, while the HI range of >35% represented early maturing accessions with high seed yields, providing a strong basis for future selection. Fluctuation in mean temperature (22.5, 22.4 and 24.7 °C) and rainfall (518, 644.6 and 287.7 mm) during the three cropping seasons (October-April) under study had a significant impact on performance of the accessions. The better average yield was observed in the third year, which might be attributed to aforementioned temperature and rainfall differences. Despite the weather impact, 10 accessions, viz., 5930, 6057, 5865, 34709, 5542, 5884, 17794, 34693, 5888, and 5944 exhibited high yield potential in all three years and are therefore recommended for lentil improvement programs in the future.