This paper proposes a new robust exponential sliding mode differentiator for estimating the (n À 1) derivatives of a non-linear function. Finite time convergence, robustness, and exactness are also ensured analytically with the proposed methodology. In addition, the results of the proposed sliding mode differentiator are extended to derive theorems for the novel state (SO) and extended state observers (ESO), which would estimate the system states as well as uncertainties recursively in finite time. Finally, three examples are implemented to validate the proposed methodologies and the obtained simulations are compared with the previously developed methods in literature.