Flexoelectricity is an electromechanical coupling phenomenon that can generate noticeable electric polarization in dielectric materials for nanoscale strain gradients. It is gaining increasing attention because of its potential applications and the fact that experimental results were initially an order of magnitude higher than initial theoretical predictions. This stimulated intense experimental and theoretical research to investigate flexoelectric coefficients in dielectric materials such as two-dimensional materials. In this study, we concentrate on the calculation of the flexoelectric coefficients of 2D-MoS2 due to a model using self-consistently determined charges and dipoles on the atoms. More specifically, we study the importance of two contributions that were neglected/omitted in previous papers using this model, namely, the charge term in the total polarization and the conservation of electric charge through a Lagrange multiplier. Our calculations demonstrate that the results for flexoelectric coefficients computed with this improved definition of polarization agree better with experimental measurements, provided that consistent definitions for signs are used. Additionally, we show how two physical contributions with opposite signs compete to give net values of flexoelectric coefficients that can be either positive or negative depending on their relative importance and give net values for the case of MoS2.