Salivary Aβ40, Aβ42, t-tau, and -tau 181 are commonly employed in Alzheimer's disease (AD) investigations. However, the collection method of these biomarkers can affect their levels. To assess the impact of saliva collection methods on biomarkers in this study, 15 healthy people were employed in the morning with six saliva collection methods. The chosen methods were then applied in 30 AD patients and 30 non-AD controls. The levels of salivary biomarkers were calculated by a specific enzyme-linked immunosorbent assay. The receiver operating characteristic was utilized to assess salivary biomarkers in AD patients. The results demonstrated that the highest levels of salivary Aβ40, Aβ42, t-tau, and -tau were in different saliva collection methods. The correlations between different saliva biomarkers in the same collection method were different. Salivary Aβ40, Aβ42, t-tau, and -tau had no significant association. Salivary Aβ42 was higher in AD than in non-AD controls. However, -tau/t-tau and Aβ42/Aβ40 had some relevance. The area under the curve for four biomarkers combined in AD diagnosis was 92.11%. An alternate saliva collection method (e.g., USS in Aβ40, UPS in Aβ42, t-tau, SSS in -tau 181) was demonstrated in this study. Moreover, combining numerous biomarkers improves AD diagnosis.