The aim of this work was to evaluate the use of a water-retaining polymer on the initial growth of coffee plants for different levels of irrigation and types of seedling. The experiment was set up in pots with a clayey soil, in a greenhouse of the Sector for Coffee Production of the Federal University of Lavras, Brazil. The experiment was arranged in a randomised block design into sub-lots with three replications, giving a total of 48 experimental units. Four levels of irrigation (25%, 50%, 75% and 100% of the available water) were randomly assigned to the lots; the type of seedling (grown in bags or tubes) were randomly assigned to the sub-lots, and the use or not of the water-retaining polymer randomly assigned to the sub-sub-lots. In the treatments with water-retaining polymer, a dilution of 1.5 Kg of polymer to 400 litres of water was used at a dose of 1.5 litres of solution per plant. The following were evaluated at intervals of 60 days during the experiment (360 days from planting): stem diameter, plant height, leaf area, number of leaves and number of plagiotropic branches on the coffee plants. Hydrated water-retaining polymer favoured the growth of coffee plants under an irrigated system. Irrigation resulted in greater plant growth, with the seedlings grown in bags showing more growth than those grown in tubes.