The excess of sodium, carbonates and bicarbonates, which are present in irrigation water, increases the pH, the electric conductivity and the percentage of interchangeable sodium in the soil, and, most of the times, it also affects crop growth. The horticultural production in greenhouses in the province of La Pampa, compared with other activities such as farming and cattle raising, is relatively new. In this thesis work, some hypothesis originated as a consequence of the chemical nature of the water and its handling were stated, bearing as the general aim to improve sustainability for the lettuce producers through the recovering of the soils degraded by irrigation with sodium bicarbonate underground waters in the petro-calcium soils of La Pampa. Physical and physic-chemical evaluations were performed in the superficial horizons of soils irrigated during different years with underground bicarbonate sodium waters. In 2 and 4-year-old tunnels with continuous localized irrigation (B1 and B2 respectively) the following treatments were applied: 1) T0, non-treated witness; 2) T1, 3.5 Mg ha-1 of granulated agricultural gypsum, 3) T2, 3.5 Mg ha-1 of agricultural gypsum + drainage, and 4) T3, drainage. The factorial design was utilized (2 and 4 years with their respective treatments) and completely at random, and subsamples were employed as measuring units. Four lettuce crop cycles were evaluated (during the months of April, August, January and May). The productivity (P), the individual weigh per plant (PIP), the electric conductivity (CE) and the pH of the soil were determined for each crop cycle. The organic material (MO), the apparent density (DA) and the interchange cations (CIC) were also determined at the beginning and end of the cycle. In the tunnel with more years of continuous irrigation all the treatments, including drainage, improved P in the four cycles in relation to T0, showing higher differences when gypsum was applied. On the contrary, B1 only showed productive response in the 2nd cycle for those treatments with gypsum application, while drainage did not have an impact in relation to T0. Considering the accumulated value of P, it was always statistically significant in tunnel B2, while in B1 it was only statistically significant until the second productive cycle. These results show that the effect achieved by the gypsum amendment was only registered until the 2nd productive cycle, provoking really abrupt drops of P and PIP 300 days after having incorporated the practice. PSI values were lower in superficial layers in treatments with amendments, while values were similar between treatments in depth. Thus, the reduction of initial PSI in the different tunnels and treatments could be corroborated. However, no significant differences were registered for the pH variable between both evaluated factors. CE varied between treatments, but no significant differences between those and the years of productions (p<0,10) could be confirmed. These results can be related to fertility rates (IMO, resulting 6,6 vs. 11,5 for B1 and B2 respectively). Drainage ditches had more impact on tunnel B2 because of the better physical conditions and permeability of the soil. These ditches were effective when soil washing was needed. During the 2nd and 3rd cycle the CE decreased abruptly and was placed below that earlier threshold.