We report a numerical study of crystallization and melting in bulk statistical homogeneous (random), homogeneous (slightly alternating), and heterogeneous (produced in a batch reaction) copolymers formed by crystallizable monomers and noncrystallizable comonomers. In our dynamic Monte Carlo simulations of lattice chains, the current model further assumes that the comonomers cannot move into crystalline regions by sliding diffusion of the chains. We find that both the overall composition and the statistical distribution of the monomers affect the phase-transition temperature, the resulting relative crystallinity, and the crystal morphology. However, the final absolute crystallinity of homogeneous copolymers seems insensitive to these parameters. Intramolecular segregation between monomers and comonomers is accompanied by crystallization, demonstrating the concept of sequence segregation or nanophase separation of statistical copolymers with assembling structures like in thermoplastic elastomers. Moreover, if crystallization of homogeneous copolymers has started but not yet completed on cooling, subsequent heating will show cold crystallization before melting, which can be attributed to insertion-mode lamellar growth. For heterogeneous copolymers, intermolecular segregation occurs on cooling before crystallization. On the basis of our observations, a pair of master melting and crystallization curves for the crystallinity of a statistical copolymer as a function of temperature are suggested to reflect the characteristic of the monomer-sequence-length distribution. This suggestion facilitates the clarification to the kinetic disturbance in local temperature regions and to the principle of some fractionation methods, such as temperature rising elution fractionation (TREF).