“…Complex modeling problems in physics, mechanics, biology, and engineering are the driving force for the theoretical and applied study of fractional calculus [2]. Since most fractional models cannot be solved analytically, many researchers resort to developing efficient and reliable numerical methods to solve fractional equations [3,4]. This paper is devoted to consider randomized estimates for fractional Carathéodory type differential equations in the following form, { 𝐶 D 𝛼 0,𝑡 𝑢(𝑡) = 𝑓(𝑡, 𝑢(𝑡)), 𝑡 ∈ [0, 𝑇], 𝛼 ∈ (0, 1) 𝑢(0) = 𝑢 0 , (1.1) where 𝑇 ∈ (0, ∞), 𝑢 ∶ [0, 𝑇] → ℝ 𝑑 , 𝑑 ∈ ℕ, and the initial condition 𝑢 0 ∈ ℝ 𝑑 .…”