Most people offer a trivial solution, for example using the FORECAST
function or other similar functions that presume a linear relationship.
But that might be wrong.
First, I suggest that you graph your historical data using Chart Wizard,
choosing the X-Y chart and finding a trendline that fits the historical data
best. If the linear trendline is truly the best fit, you might use
FORECAST.
Second, you should reflect on patterns of sales that might be
circumstantial, not numerical. For example, seasonal variations, or
interpendencies with the economic trends of other industries. Integrating
that into your forecast model goes beyond "how to" questions about Excel
formulas.
This only scratches the surface. But it might give you something to think
about before jumping into writing formulas.