First, there is some major problem with your question. Notice that there is no chance to write
$$f^*(w^*)=\sum_{something-I-dont-see} f^*_{?}(w^*)\otimes f^*_{?}(w^*),$$
because the left hand side is linear in $w^*$ and the right hand side is quadratic!
But something can be done!
You said that $(U\otimes V)^*\simeq U^* \otimes V^*.$ However, it is only true, because $U$ and $V$ have finite dimensions. In general, for $U,V$ with arbitrary dimensions we only have an injective map $\phi:U^* \otimes V^*\to(U\otimes V)^*$ given by formula
$$\phi\left(\sum_i\alpha_i\otimes\beta_i\right)\left(\sum_jx_j\otimes y_j\right)=\sum_{i,j}\alpha_i(x_j)\beta_i(y_j).$$
Here you have a related discussion about this problem.
So your question can be reformulated as
Problem. Let $f:U\otimes V\to W$ be given and let $f^*:W^*\to(U\otimes V)^*$ denote the dual map. Is there a map $\bar{f}:W^*\to U^*\otimes V^*$ such that
$$f^*=\phi\circ \bar{f}?$$
Answer. If $U,V$ have finite dimension, then there is such $\bar{f}$ but not defined in canonical manner (it depends on basis of $U$ and $V$).
In fact, if $u_1,\dots,u_n$ and $v_1,\dots,v_m$ are basis in $U$ and $V$ respectively, then we have a dual basis $\omega_1,\dots,\omega_n$ and $\eta_1,\dots,\eta_m$ in $U^*$ and $V^*$ respectively. As a result $\lbrace \omega_i\otimes\eta_j\rbrace_{i,j}$ form a base in $U^*\otimes V^*$ and $\lbrace \phi(\omega_i\otimes\eta_j)\rbrace_{i,j}$ form a base in $(U\otimes V)^*$. Thus, for any $w^*\in W^*$ we have that
$$f^*(w^*)=\sum_{i,j}a_{i,j}(w^*)\phi(\omega_i\otimes\eta_j).$$
Putting $\bar{f}=\phi^{-1}\circ f^*$ we obtain that
$$\bar{f}(w^*)=\sum_{i,j}a_{i,j}(w^*)\omega_i\otimes\eta_j.$$
This is all I can tell for certain! Notice also that we do not need $W$ to have finite dimension.