Often industrial variables can be difficult to measure due to such factors as extreme conditions or complex compositions. In such cases, soft sensors have been developed that use available system information and measurements to estimate these difficult-to-obtain variables. In practice, the measurements that are to be estimated by a soft sensor are often infrequently measured or delayed. Occasionally, these sampling times or delays are time varying. At present, most research has considered these parameters to be time invariant, and thus, there is a need to consider the time-varying case. Therefore, this paper will evaluate the impact of time-varying delays and sampling times for the design of a data-driven soft sensor. Modifications will be proposed that will increase the robustness and performance of the soft sensor. The reliability of the estimate will be shown using the Bauer- Premaratne-Duran Theorem. Furthermore, the proposed soft sensor system will be tested using simulations of a continuous stirred tank reactor (CSTR) and an reverse osmosis plant. Simulation showed that the modified soft sensor gives good estimates, whereas the traditional soft sensor gives an unstable estimate for the CSTR and reverse osmosis plant. (C) 2020 Elsevier Ltd. All rights reserved.