<p>The present study was aimed to use multiple linear regression (MLR) to predict the most sensitive quantitative and qualitative traits influencing yield per plant, downy mildew resistance, and artificial neural network (ANN) to predict traits influencing nutraceutical content (ascorbic acid and β-carotene) in cucumber. Twenty-nine genotypes were evaluated across two years under open field conditions using a randomized block design with three replications. MLR models identified fruit weight, number of fruits per plant, and fruit length as key quantitative traits affecting yield per plant (adjusted R² = 0.834). Orange fruit color at maturity, dense ovary pubescence, and medium leaf blade margin dentation were significant qualitative traits (adjusted R² = 0.59). For downy mildew resistance, leaf blade length, fruit weight, number of fruits per plant, ascorbic acid and total sugar contents were the most predictive quantitative characters (adjusted R² = 0.59). Morphologically, cylindrical fruit shape and creamy white color affected PDI (adjusted R² = 0.52). ANN models showed strong generalization, especially for β-carotene prediction (testing error 0.333). Peduncle length, fruit girth, and orange fruit color at maturity were important for nutraceutical content. Based on D² analysis, 8 genotypes were crossed in 8 × 8 half diallel design without reciprocals. Genetic control of the quantitative characters revealed overwhelming importance of non-additive gene action indicating deferred selection or hybrid breeding. Genetic control across nine morphological characters using χ² goodness-of-fit test indicated preponderance of digenic gene action with presence of epistasis. A comprehensive understanding of these genetic interactions is essential for improving cucumber lines in terms of yield, quality and disease resistance.</p>