<p id="p00010">Self-esteem, as a core individual variable, has a long-standing inverse relationship with materialism. However, some studies have found that people with high self-esteem have higher materialistic orientation than those with low self-esteem, which contradicts previous research. In this connection, researchers need to rethink the relationship between these two variables. The heterogeneity hypothesis of high self-esteem holds that there are secure high self-esteem and fragile high self-esteem in high self-esteem groups, which may guide our understanding of this contradictory relationship. Therefore, the purpose of this study was to explore whether materialistic orientation was associated with different sub-types of high self-esteem, that is, that high self-esteem is heterogeneous, and to explore whether the relationship would change under self-threat situations. <p id="p00015">The results of study 1 indicated that the relationship between high fragile self-esteem and materialism was positive, while the relationship between high secure self-esteem and materialism was negative. The materialism of people with high fragile self-esteem was significantly higher than those with high secure self-esteem. The results of study 2 suggested that attractiveness threat moderated the relationship between heterogeneous high self-esteem and materialism. Specifically, in the attractive threat scenarios, materialism of secure high self-esteem was low, while materialism of fragile high self-esteem was high. In the non-attractive threat situation, the difference in materialism measures between them was insignificant. This finding was further replicated in study 3 where intellectual threat moderated the relationship between heterogeneous high self-esteem and implicit materialism. Notably, in the context of an intellectual threat, implicit materialism of the secure high self-esteem group was low, while implicit materialism of the fragile high self-esteem group was high. In the non-intellectual threat situation, the difference in implicit materialism scores was not significant. <p id="p00020">Therefore, from the perspective of heterogeneous high self-esteem, this study essentially proved the low materialism level of those with secure high self-esteem and high materialism level of those with fragile high self-esteem. It signifies that the two conclusions of previous studies are reasonable. If there exists a greater number of secure high self-esteem individuals among a population of high self-esteem persons, the relationship between high self-esteem and materialism is negative. On the contrary, if there is a greater number of fragile high self-esteem individuals, the relationship is positive, thereby solving the puzzle of the contradictory relationship between high self-esteem and materialism. Furthermore, this study can urge educators to pay more attention to the cultivation of children's secure high self-esteem rather than unquestioningly cultivating high self-esteem. <p id="p00025">To delve deeper into the impact of high self-esteem on materialism orientation, three studies were conducted. In study 1, 420 college students from 4 universities were surveyed by questionnaire to explore the correlation between heterogeneous high self-esteem and materialism. Study 2 was aimed to verify the causal relationship, and used attractiveness threat to self-esteem to investigate whether the attractiveness threat played a moderating role in the relationship between different types of high self-esteem and materialism where participants with fragile high self-esteem and secure high self-esteem were randomly assigned either to the attractiveness threat group or the control group. They were asked to compare their own attractiveness to the most and the least attractive images of the same sex selected from pre-test, respectively. Study 3 used intellectual threat as a form of self-threat priming and measured materialism at the implicit level where heterogeneous high self-esteem participants were randomly assigned to the intellectual threat group or the control group. The former completed 13 difficult questions from Raven's Progressive Matrices and received negative feedback, while the latter completed 13 simple questions from Raven's Progressive Matrices and received positive feedback.