BACKGROUND: The association between telomere length (TL) and autism spectrum disorder (ASD) has received much attention. However, previous observational studies have yielded inconclusive evidence regarding this relationship. Our study aims to elucidate the causal relationship between TL and ASD using bidirectional Mendelian randomization (MR). METHODS: We employed the largest genome-wide association studies (GWAS) summary statistics for TL (sample size = 472,174) and ASD (sample size = 46,351). The primary MR analysis method was the inverse-variance weighted (IVW) method, complemented by the MR-Egger method, weighted median (WM) method, and MR-PRESSO. Additionally, sensitivity analyses including Cochran's Q test, the intercept of MR-Egger regression, the global test of MR-PRESSO, and the leave-one-out analysis were conducted in our study. RESULTS: The primary MR analysis indicated a significant association between ASD and shorter TL (IVW: OR = 0.98, 95% CI: 0.96-0.99, p = 0.03). However, no significant association was found in the reverse direction MR analysis (IVW: OR = 1.06, 95% CI: 0.94-1.23, p = 0.35). Raw and outlier-corrected MR estimates from MR-PRESSO were consistent with the IVW results. Sensitivity analyses confirmed the robustness of these findings. CONCLUSIONS: Our study indicated that individuals with ASD have shorter TL, however, shorter TL does not appear to increase the risk of ASD.