Decrease the sample size. Other things being equal, a smaller sample will have a larger confidence interval (and a larger sample will have a smaller confidence interval).
Increase the proportion of the confidence interval. Confidence intervals are designed to cover a certain proportion of the population. Popular choices are 95% and 99%, but any percent can be used, from 1 to 99.99. Other things being equal, a 99% confidence interval will be larger than a 95% confidence interval.
Get poorer data. The confidence interval is also affected by the quality of your data. Worse data (that is, data that is less reliable or less accurately measured) will give larger confidence intervals than better data.