The atherogenic index of plasma (AIP), a novel composite lipid index, is closely linked to cardiovascular disease (CVD). However, lipid levels fluctuate dynamically, and it is unclear whether there are differences in the association of single-timescale, multiple-timescale, or AIP change trajectories with new-onset cardiovascular disease. Hence, the aim of this study was to investigate the correlation between different AIP parameters and the occurrence of CVD. Data were derived from the China Health and Retirement Longitudinal Study (CHARLS) conducted in 2011, 2015, 2018, and 2020, focusing on middle-aged and elderly populations aged over 45 years. Changes in AIP were classified into three groups using K-means cluster analysis: the low-level growth group (Class 1), the medium-level growth group (Class 2), and the high-level decline group (Class 3). Furthermore, participants were grouped based on tertiles (T) of cumulative AIP (Cum-AIP). Our multivariate logistic regression model integrated adjustments for potential confounders in order to investigate the association between Cum-AIP and the occurrence of CVD. Additionally, we employed restricted cubic spline (RCS) modeling to illustrate the dose-response relationship of baseline AIP, mean AIP, and Cum-AIP with CVD risk. During the 5-year follow-up period, 927 participants experienced the onset of CVD. After controlling for various potential confounding factors, it was observed that individuals in Class 2 demonstrated a notably heightened risk of CVD (OR = 1.23, 95% CI: 1.03, 1.46) and stroke (OR = 1.35, 95% CI: 1.02, 1.80) in comparison to those in Class 1. However, there was no significant difference in the risk of heart disease (OR = 1.21, 95% CI: 0.99, 1.48). In contrast, a noteworthy correlation was solely observed in the Class 3 group concerning the risk of stroke occurrence (OR = 1.60, 95% CI: 1.06, 2.42). The adjusted OR (95% CI) for CVD in the T2 and T3 groups were 1.21 (1.00, 1.46) and 1.30 (1.05, 1.62), respectively, compared to the T1 Cum-AIP group (P for trend = 0.017). Through the RCS model, we identified a positive and linear relationship between baseline AIP, mean AIP, and Cum-AIP with the incidence of CVD. However, the association between baseline AIP and CVD was weak. Sustained elevation of AIP is linked to a heightened risk of CVD in the general population. The elevated mean, and Cum-AIP levels are associated with a heightened risk of CVD. These findings indicate that AIP can serve as a valuable indicator of dyslipidemia, and continuous monitoring and early intervention targeting AIP may contribute to a further reduction in the incidence of CVD.