I am trying to prove using similar triangles without considering an arbitrary point on the diagonals which is the most common proof what we see in text books.
Let $HI=a$, $IJ=b$, $JC=c$ ,$CH=d$, $HL=x$, $LJ=e-x$, $CL=f-y$ and $LI=y$
we have $$\Delta HLC \sim \Delta ILJ $$ so we have
$$\frac{d}{b}=\frac{f-y}{e-x}=\frac{x}{y} \tag{1}$$ Also
$$\Delta CLJ \sim \Delta HLI$$ so
$$\frac{a}{c}=\frac{e-x}{y}=\frac{f-y}{x} \tag{2}$$ from $(1)$ and $(2)$ is it possible to prove that $$ac+bd=ef$$




Not directly, as far as I can see. For one thing, Ptolemy's theorem "decays" nicely to $a c = a c$ in the degenerate case where $I \equiv J, b = 0, e = a, f = c$, while similarity-based proofs would not directly translate to the trivial case.
That said, the similarity of both pairs of triangles is equivalent to calculating the Power of the Point $L$ with respect to the given circle $LH\;LJ = LI\; LC$. The power of the point is related to the Inversion transformation, and it is indeed possible to prove Ptolemy by Inversion. As said, however, that's not exactly an immediate consequence of the given similarity.