We consider the determination of an unknown spatial-dependent coefficient for the pseudoparabolic system with Dirichlet boundary condition from nonlocal inversion input data. Based on the positive property of the solution to the direct problem, the uniqueness and the Lipschitz conditional stability of this nonlinear inverse problem are addressed in Hölder space, by the principle of contraction mapping for specially chosen weight function describing the nonlocal inversion input. Then, an iterative algorithm is established in terms of the fixed point equation for the unknown potential to construct the approximate solution to this inverse problem, with rigorous convergence analysis and the error estimate on the iterative sequence, which also leads to the optimal approximation error for choosing the stopping rule of the iteration process. The proposed iterative algorithm is compared with classical optimization scheme with regularizing term, showing the advantages of our scheme. Some numerical results are presented to validate our proposed scheme. [ABSTRACT FROM AUTHOR]