© 2020 Elsevier GmbHUltrasonic probe extraction of phenolic compounds from olive leaves was modelled and optimized by using soft computing methods. The effects of amplitude, irradiation time, solvent (MeOH/EtOH) concentration, probe position and duty cycle on total phenolic content (TPC), oleuropein content (OLEc) and total antioxidant activity (TAA) of olive leaf extracts were investigated in multi-objective perspective. The predicted quadratic linear models were obtained by using the least square (LS) approach. They were optimized through the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and Pareto sets were obtained. A compromise solution was chosen among many non-dominated solutions of experimental conditions by using multi-criteria decision making (MCDM) methods. The fuzzy c-means clustering algorithm (FCM) method was used to select the compromise solution. The optimum conditions of MeOH/H2O ultrasonic probe extraction were determined as; amplitude 81.91 %, irradiation time 14.22 min, MeOH concentration 76.97 %, probe position 3.89 cm, duty cycle 0.93 % and the experimental results obtained under these conditions were 7.19 ± 0.44 mg OLE/g dw, 5.95 ± 0.19 mg/g dw, 88.36 ± 2.47 % for TPC, OLEc and TAA, respectively. For EtOH/H2O ultrasonic probe extraction; the observed results were 5.24 ± 0.08 mg OLE/g dw, 2.22 ± 0.08 mg/g dw, 80.16 ± 1.25 % for TPC, OLEc and TAA, respectively at the optimum conditions of amplitude 79.16 %, irradiation time 12.90 min, EtOH concentration 73.40 %, probe position 1.59 cm, duty cycle 0.97 %.