Uncertain water supplies resulting from changing climatic conditions in western North America led to this investigation of the role of crop load reduction in maintaining performance of high-density ‘Ambrosia’ apple (Malus ×domestica) on M.9 rootstock. A split-plot experimental design was imposed for three growing seasons (2007–09) with six replicates of four main plot irrigation treatments and three crop load subplots comprised of three trees. Four season-long irrigation (Irr) treatments were applied through 2 × 4 L·h−1 drip emitters per tree and included Irr1) control [100% evapotranspiration (ET) replacement], Irr2) 50% ET replacement, Irr3) 50% ET replacement to half the emitters, and Irr4) an increasingly severe treatment commencing at 50% ET replacement (once every 2 days) in 2007 and progressing to 25% and 18% ET replacement, 2008–09. Three target crop loads were established annually, 4–5 weeks after bloom as low (2.5, 3, and 3.75), medium (4.5, 6, and 7.5), and high (9, 12, and 15) fruit/cm2 trunk cross-sectional area (TCSA) 2007–09, respectively, by hand thinning around 4 weeks after bloom. Volumetric soil moisture contents generally reflected the amount of water applied and ranged from 20% for control (Irr1) to <10% for Irr4. Both irrigation and crop load treatments affected midday stem water potential more than leaf photosynthesis and stomatal conductance (gS). By the 2nd and 3rd year stem potential values for irrigation treatments ranged from a maximum of −1.0 to −1.3 MPa for Irr1 to minimums ≤-2.0 MPa for Irr4. gS decreased as midday stem potential decreased, but at any given stem potential value was greater at high crop loads, presumably in response to an increased demand for photosynthates. Fruit size decreased as crop load increased, but as irrigation deficits became more severe, fruit size was more closely correlated with stem water potential than gS. Consequently, fruit size was controlled by two mechanisms, competition for photosynthates and the effects of plant water status on gS. Negative linear relationships between crop load and average fruit size were used to determine the crop load required to produce an average fruit size of 200 g at different irrigation deficits. It was not possible to achieve adequate fruit size when applications were very low, as at 18% to 25% ET in Irr4. Crop load reduction around mid-June had no negative consequences for fruit quality, enhancing fruit color, and soluble solids concentration (SSC) and did not affect the incidence of sunburn, internal breakdown or bitter pit at harvest.