The current work presents a distributed estimation approach with a topology-switching structure and introduces an adaptive self-triggered strategy (ASTS) to minimize energy consumption during inter-node communication. In the filter design, the network's communication topology is modeled as a time-varying process, with switching governed by a homogeneous Markov chain and a probabilistic transition matrix containing partially unknown data. Filter design feasibility is verified using Lyapunov stability theory and linear matrix inequality (LMI) method, which are used to determine the filter parameters. Numerical simulation and practical experiment with a continuous stirred tank reactor validate the proposed approach.