Via the theory of hilbert polynomials, I think the following is true:
Suppose $X$ is a closed subscheme of $P^n$ of degree $d$, and $Y$ is also such one but is also a complete intersection and also doesn't share any component with $X$, then (assuming the the sum of dimensions is at least $n$)
$$deg(X \cap Y) = deg(X) \cdot deg(Y)$$
To prove this, it suffices to show this for $Y$ a hypersurface $f$, which is proved via the exact sequence
$$ S_*(X) \to ^f S_* (X) \to S_*(X)/(f)$$
Is this true if $X,Y$ are not complete intersection?
Jake brought an example that illustrates the problems that occur when $Y$ is a complete intersection in a bad way. Let me show that everything goes wrong when both are not complete intersections.
In $P^4$, a union of two generic planes is not a complete intersection (note they meet at a point). Take two pairs (of union of two planes) that all meet the same point. I claim they do the job. Each is certainly of degree $2$, so you'd expect the intersection to be of degree $4$. But, computing the Hilbert polynomial; one easily finds it's $5$ instead!