Ring oscillator-based true random number generators (RO-TRNGs) are widely used to generate unpredictable random numbers for cryptographic systems.Entropy is usually adopted to quantitatively measure the unpredictability of a TRNG.There have been several stochastic models such as the time-oriented and phase-oriented ones built to evaluate the entropy of Elementary-RO-TRNGs with single oscillator.However,these models are not suitable for the TRNGs composed of multiple oscillators (Multiple-RO-TRNGs),which can obtain more randomness and higher throughput.Considering this,we propose a sequence-oriented stochastic model for the entropy evaluation of RO-TRNGs,named the first-order stationary Markov source model.This model is extensible for the Multiple-RO-TRNGs.Based on that,we present a detailed method to determine the entropy of Multiple-RO-TRNGs.Our proposed model is verified by experiments.Besides,our method can also be a guide to design RO-TRNGs with both high entropy and high throughput.