Associating environmental stresses (ESs) with built-in test (BIT) output is an important means to help diagnose intermittent faults (IFs). Aiming at low efficiency in association of traditional time stress measurement device (TSMD), an association model is built. Thereafter, a novel approach is given to evaluate the integrated environmental stress (IES) level. Firstly, the selection principle and approach of main environmental stresses (MESs) and key characteristic parameters (KCPs) are presented based on fault mode, mechanism, and ESs analysis (FMMEA). Secondly, reference stress events (RSEs) are constructed by dividing IES into three stress levels according to its impact on faults; and then the association model between integrated environmental stress event (IESE) and BIT output is built. Thirdly, an interval grey association approach to evaluate IES level is proposed due to the interval number of IES value. Consequently, the association output can be obtained as well. Finally, a case study is presented to demonstrate the proposed approach. Results show the proposed model and approach are effective and feasible. This approach can be used to guide ESs measure, record, and association. It is well suited for on-line assistant diagnosis of faults, especially IFs. Associating environmental stresses (ESs) with built-in test (BIT) output is an important means to help diagnose intermittent faults (IFs). Aiming at low efficiency in association of traditional time stress measurement device (TSMD), an association model is built. First, the selection principle and approach of main environmental stresses (MESs) and key characteristic parameters (KCPs) are presented based on fault mode, mechanism, and ESs Analysis of secondary stress (RSEs) is constructed by dividing IES into three stress levels according to its impact on faults; and then the association model between integrated environmental stress events (IESE) and BIT output is built. Thirdly, an interval gray association approach to evaluate IES level is proposed due to the interval number of IES value. Finally, a cas Results show the proposed model and approach are effective and feasible. This approach can be used to guide ESs measure, record, and association. It is well suited for on-line assistant diagnosis of faults, especially IFs.