<p>In real-world applications, the estimation of finite population variance becomes difficult when the underlying variability is hard to manage or control. The use of auxiliary information greatly enhances the precision and efficiency of estimators for unknown population parameters. The objective of this article is to introduce a novel family of ln cum-exponential-type estimators designed for estimating the finite population variance of the study variable in stratified random sampling. The term ‘ln cum-exponential-type estimator’ refers to a class of estimators that combines the natural logarithmic and exponential transformations of auxiliary information within a unified estimation framework. We determine the bias and mean squared error (MSE) of the proposed estimator by applying first-order approximation techniques. Empirical and simulation analyses are carried out to assess the performance and real-world feasibility of the proposed estimator. Across both the empirical studies and all simulated conditions, the introduced estimators consistently achieve lower MSE and higher Percentage relative efficiency (PRE) values as compared to the conventional estimators. The increased efficiency and reliability of proposed estimator emphasize its usefulness for real-world data applications.</p>