Considering the research model involved latent variables that could not observed directly, structural equation modeling (SEM) with component-based approach will applied with SmartPLS 2.0 from [15] to estimate the parameters. This method was chosen with following considerations: 1. As one method in soft modeling group, component-based SEM is not too strict about fulfilling the assumptions of normality, as required under covariance-based SEM (AMOS, LISREL); 2. The existence of formative and reflective indicators that cannot be handled by covariance-based SEM [16], and; 3. Can be used without any loss of power of the test for very small sample. Component-based SEM has proven its use in sample size is only 6 [17]. In studying the relationship between people participation, social capital, and ecotourism sustainability at Kintamani District in Bali; we conducted two types of analysis i.e. measurement model analysis to evaluate the appropriateness of items/indicators in explaining latent variable, and inner or structural model analysis to measure effect of exogenous towards endogenous latent variable and test the hypotheses.