In carbon ion beams, biological effects vary along the ion track; hence, to quantify them, specific radiobiological models are needed. One of them, the local effect model (LEM), in particular version I (LEM I), is implemented in treatment planning systems (TPS) clinically used in European particle therapy centers. From the physical properties of the specific ion radiation, the LEM calculates the survival probabilities of the cell or tissue type under study, provided that some determinant input parameters are initially defined. Mathematical models can be used to predict, for instance, the tumor control probability (TCP), and then evaluate treatment outcomes. This work studies the influence of the LEM I input parameters on the TCP predictions in the specific case of prostate cancer. Several published input parameters and their combinations were tested. Their influence on the dose distributions calculated for a water phantom and for a patient geometry was evaluated using the TPS TRiP98. Changing input parameters induced clinically significant modifications of the mean dose (up to a factor of 3.5), spatial dose distribution, and TCP predictions (up to factor of 2.6 for D50). TCP predictions were found to be more sensitive to the parameter threshold dose (Dt) than to the biological parameters α and β. Additionally, an analytical expression was derived for correlating α, β and Dt, and this has emphasized the importance of [Formula: see text]. The improvement of radiobiological models for particle TPS will only be achieved when more patient outcome data with well-defined patient groups, fractionation schemes and well-defined end-points are available.