OBJECTIVE The decline in C-peptide in the 5 years after diagnosis of type 1 diabetes has been well studied, but little is known about the longer-term trajectory. We aimed to examine the association between log-transformed C-peptide levels and the duration of diabetes up to 40 years after diagnosis. RESEARCH DESIGN AND METHODS We assessed the pattern of association between urinary C-peptide/creatinine ratio (UCPCR) and duration of diabetes in cross-sectional data from 1,549 individuals with type 1 diabetes using nonlinear regression approaches. Findings were replicated in longitudinal follow-up data for both UCPCR (n = 161 individuals, 326 observations) and plasma C-peptide (n = 93 individuals, 473 observations). RESULTS We identified two clear phases of C-peptide decline: an initial exponential fall over 7 years (47% decrease/year [95% CI −51, −43]) followed by a stable period thereafter (+0.07%/year [−1.3, +1.5]). The two phases had similar durations and slopes in patients above and below the median age at diagnosis (10.8 years), although levels were lower in the younger patients irrespective of duration. Patterns were consistent in both longitudinal UCPCR (n = 162; ≤7 years duration: −48%/year [−55, −38]; >7 years duration −0.1% [−4.1, +3.9]) and plasma C-peptide (n = 93; >7 years duration only: −2.6% [−6.7, +1.5]). CONCLUSIONS These data support two clear phases of C-peptide decline: an initial exponential fall over a 7-year period, followed by a prolonged stabilization where C-peptide levels no longer decline. Understanding the pathophysiological and immunological differences between these two phases will give crucial insights into understanding β-cell survival.