In this study, an attempt has been made to explore the influence of four inflow boundary conditions on solute plume migration and compared their concentrations. A numerical model was used to simulate the solute transport for different boundary conditions at the injection well with three different recharge and six hydraulic conductivity scenarios. Dirichlet concentration boundary, constant point source, and two Robin boundary conditions were applied to exemplify their effects on the outcomes. Significant variations were visible in the solute concentration profiles and their respective spreading patterns. Results show discrepancies between solutions obtained from the first-type and the third-type inflow boundary conditions for smaller Peclet numbers. The applicability of the M5′ model tree, a tree-based machine learning approach, was investigated and thereby exploited its capability to interpret the plausible functional dependency among the input parameters. The model tree also produced a dominancy structure within the parameter space with its combined classification and regression features. It was concluded that adopting a general boundary condition or generalization of solutions remained highly challenging, given the different input space parameters dictating a hydrogeological model. © 2021, The Author(s), under exclusive licence to Springer Nature Switzerland AG.