Judging by the screenshots you've posted, the culprit is the ill-advised use of \left(
and \right)
in the inline math formula. Material inside a \left( ... \right)
construct is never broken across lines. Never, ever. If you need to enlarge the opening and closing parentheses of the inline formula, use \Bigl(
and \Bigr)
instead.
As proof, consider the following screenshot and associated LaTeX code. It shows that zero line break possibilities are available if \left( ... \right)
is used in the formula. In contrast, TeX finds four such possibilities if \Bigl( ... \Bigr)
is used. I leave it to you to decide which method you should be using.

\documentclass{article}
\usepackage{mathpazo} % looks like you're using a Palatino font
\setlength\textwidth{1mm} % choose an extremely narrow measure
\setlength\parindent{0pt} % just for this example
\begin{document}
\mbox{\emph{With} \texttt{\string\left} and \texttt{\string\right}:
\textbf{zero} linebreak possibilities in the formula}
\medskip
$\left(\omega=\frac{d\theta}{dt}=
\int_0^{\dot{\theta}} d(\dot{\theta});\allowbreak
\omega^2=\int_{0}^{\dot{\theta}}d(\dot{\theta}^2) \right)$
\bigskip
\mbox{\emph{Without} \texttt{\string\left} and \texttt{\string\right}:
\textbf{four} [!] linebreak possibilities}
\medskip
$\Bigl(\omega=\frac{d\theta}{dt}=
\int_0^{\dot{\theta}} d(\dot{\theta});\allowbreak
\omega^2=\int_{0}^{\dot{\theta}}d(\dot{\theta}^2) \Bigr)$
\end{document}