Regulated deficit irrigation (RDI) involves water stress management in different phenological periods throughout the season. Research in olive trees (oil production) suggested RDI during pit hardening based in predawn and midday stem water potential (SWP) thresholds. However, the previous thresholds may not be extrapolated to table olive because fruit size, a very important feature in the table olive yield quality, is very sensitive to water stress. RDI in table olive deserve further research to determine the optimal water potential thresholds and the duration of the RDI periods for the specificity of the crop (low crop load to promote high fruit size). The aim of this work was to study different RDI schedules during pit hardening, considering different levels and durations of water stress. The experiment was performed in the 2015, 2016 and 2017 seasons, in a commercial mature table olive orchard (cv. Manzanilla) in Dos Hermanas (Seville, Spain). Control treatments were based on midday SWP measurement in order to optimize the water status with values around −1.4 MPa. Two RDI treatments were applied during pit hardening, dated (according to the changes in longitudinal fruit growth) from mid-June to the last week of August) to maintain water potential values around −2 MPa (RDI-1) and −3.5 MPa (RDI-3). Another RDI treatment (RDI-2) received irrigation to maintain values around −3.5 MPa but the recovery was performed at early July in order to obtain different durations of water stress. Irrigation strategies were evaluated with water relations measurements (soil moisture, gas exchange), fruit and shoot growth and quality and quantity yield indicators. Yield was not significantly affected in any of the RDI treatments with an ANOVA analysis. However, fruit drop estimated as the percentage of fruit lost only in the period of water deficit was related with water stress parameters (SWP and stress integral, IS). In addition, the relationship between fruits size and these latter parameters were significant and change according to yield level. Irrigation treatments did not affect next season yield because shoot growth and number of inflorescence at the beginning of each season were not different. RDI effect changed according to yield level, mainly in relation with fruit size. Data suggest that yield levels up to 12 t ha−1 were possible to manage RDI without affecting fruit size or reducing commercial quality