(1) Background: Crop yields in China’s arid and semi-arid regions are limited by water shortages. Exploring the interactions and resource utilization among agroforestry species is key to maintaining diversified agricultural production. (2) Objective: An apple–watermelon agroforestry system and watermelon sole-cropping system were compared to quantify how resource availability (light, water) and watermelon performance (leaf photosynthetic rate, growth, and yield) change with irrigation strategies. (3) Methods: A three-year apple and watermelon field experiment was conducted in a young apple orchard in the arid area of central Ningxia to test the effect of light competition and irrigation systems on light environment, leaf photosynthetic rate, plant growth, and yield in watermelon. The experiment encompassed two planting patterns: (i) apple–watermelon agroforestry (AF) and watermelon sole-cropping (SC) and (ii) three irrigation quotas (W1: 105 mm, W2: 210 mm, and W3: 315 mm). (4) Results: The results show that the agroforestry planting pattern extended the growth period of watermelon and increased the leaf area index. Mean daily shade intensity increased by 16.02% from 2020 to 2022. The land equivalent ratio (LER) was > 1 in 2021 and 2022. The SWC, leaf photosynthetic rate, LAI, and yield of watermelon in an agroforestry planting pattern were lower than when in a sole-cropping planting pattern. However, under the W1 irrigation strategy, the total soluble solids of the agroforestry planting pattern were 2.27% higher than those of the sole-cropping pattern, and the yield of the agroforestry planting pattern was 2.59% higher than that of the sole-cropping pattern. Under the W3 irrigation strategy, the average watermelon weight in the agroforestry planting pattern was 2.85% higher than that of the sole-cropping pattern. A path analysis showed that the agroforestry planting pattern can increase the yield by increasing soil water content, which is different from the sole-cropping pattern. (5) Conclusions: The results confirm that the apple–watermelon agroforestry planting pattern reduced watermelon yields. However, the LER of the agroforestry system was greater than 1.0. It is reasonable to plant watermelons in young apple forests.