This link introduces variable chromosome lengths (VCL) in the context of a genetic algorithm (GA). This concept is implemented on structural topology optimization but is also suitable for a broader class of design problems.
If you already have variable length chromosomes, then it shouldn't matter how you do it, you just need to select a crossover point for each of them, and then crossover as normal.
For example, on your chromosomes, I have selected two points (.) at random:
1 6 8 9.0 3 4 7 5
3 6.5 7 8 5
With resulting chromosomes:
1 6 8 9.5 7 8 5
3 6.0 3 4 7 5
If you wish to learn the Genetic Algorithm and Genetic Programming then visit this Artificial Intelligence Course.