If I can remember well, Landau and Lifschitz first introduce the probabilistic interpretation of the wave function, being
$$\rho(\mathbf{r},t) = \psi(\mathbf{r},t)^* \psi(\mathbf{r},t) \ ,$$
so that normalization condition holds
$$1 = \int_{\mathbf{r} \in V} \psi(\mathbf{r},t)^* \psi(\mathbf{r},t) = \langle\psi(t) | \psi(t) \rangle $$
the probability density of the coordinates.
Now,
- if we trust the superposition observed in nature the equation must be linear, and thus the equation must be of the form
$$\dfrac{d}{dt}{|\psi \rangle} = \hat{L} |\psi \rangle$$
if we trust (since there is an experimental evidence) the probabilistic interpretation of the wave function, and the fact that normalization holds at every time $t$, the operator of the linear equation must have the form $\hat{L} = -i c \hat{H}$, being $c = \frac{1}{\hbar}$ a constant involving Planck's constant, and $\hat{H}$ is the Hamiltonian operator, in order to match the experimental results. You can recover classical mechanics as well, as an example looking for Ehrenfest theorem.
The solution of the Schrodinger equation
$$|\dot{\psi} \rangle = -\frac{i}{\hbar}\hat{H} | \psi \rangle$$
reads (if the Hamiltonian is independent from time)
$$|\psi (t) \rangle = e^{-\frac{i}{\hbar} \hat{H} t} | \psi(0) \rangle \ ,$$
being $ e^{-\frac{i}{\hbar} \hat{H} t}$ the unitary evolution operator, and the normalization condition always satisfied, if it holds at the initial time,
$$\langle \psi(t) | \psi(t) \rangle = \langle e^{-\frac{i}{\hbar} \hat{H} t}\psi(0) | e^{-\frac{i}{\hbar} \hat{H} t} \psi(0) \rangle = \langle \psi(0) | \underbrace{e^{\frac{i}{\hbar} \hat{H} t} e^{-\frac{i}{\hbar} \hat{H} t}}_{=1} \psi(0) \rangle = \langle \psi(0) | \psi(0) \rangle= 1 \ .$$
Remarks. In the latter expression, the fact that the Hamiltonian operator is self-adjoint has been used.
The unitary operator can be interpreted with a series expansion,
$$e^{-\frac{i}{\hbar} \hat{H}t} = \sum_{k=0}^{\infty} \frac{\left(-\frac{i}{\hbar} \hat{H} \right)^k}{k!} \ ,$$
and written in a "diagonal form" in the base of stationary states (here meant to be discrete), i.e. the eigenvectors of the Hamiltonian as
$$\begin{aligned} e^{-\frac{i}{\hbar} \hat{H}t} & = \underbrace{|\psi_a\rangle \langle \psi_a}_{\hat{I}} | e^{-\frac{i}{\hbar} \hat{H}t} \underbrace{|\psi_b\rangle \langle \psi_b |}_{=\hat{I}} = \\
& = \psi_a\rangle \sum_{k=0}^{\infty} \langle \psi_a | \sum_{k=0}^{\infty} \frac{\left(-\frac{i}{\hbar} \hat{H} \right)^k}{k!} |\psi_b\rangle \langle \psi_b | = \\
& = \psi_a\rangle \sum_{k=0}^{\infty} \underbrace{\frac{\left(-\frac{i}{\hbar}E_b\right)^k}{k!}}_{e^{-i \frac{E_b}{\hbar}t}} \underbrace{\langle \psi_a |\psi_b\rangle}_{=\delta_{ab}} \langle \psi_b | = \\
& = | \psi_a \rangle \langle \psi_a | e^{-\frac{i}{\hbar}E_a t} \ .
\end{aligned}$$